Color Classification and Texture Recognition System of Solid Wood Panels

Article
Color Classification and Texture Recognition System of Solid
Wood Panels
Zhengguang Wang † , Zilong Zhuang † , Ying Liu *, Fenglong Ding and Min Tang

 College of Mechanical and Electronic Engineering, Nanjing Forestry University, Nanjing 210037, China;
 nanlinwzg@njfu.edu.cn (Z.W.); zzl0702@njfu.edu.cn (Z.Z.); dfl@njfu.edu.cn (F.D.); tangmin@njfu.edu.cn (M.T.)
 * Correspondence: liuying@njfu.edu.cn
 † These authors contributed equally to this work.

 Abstract: Solid wood panels are widely used in the wood flooring and furniture industries, and
 paneling is an excellent material for indoor decoration. The classification of colors helps to improve
 the appearance of wood products assembled from multiple panels due to the differences in surface
 colors of solid wood panels. Traditional wood surface color classification mainly depends on workers’
 visual observations, and manual color classification is prone to visual fatigue and quality instability.
 In order to reduce labor costs of sorting and to improve production efficiency, in this study, we
 introduced machine vision technology and an unsupervised learning technique. First-order color
 moments, second-order color moments, and color histogram peaks were selected to extract feature
 vectors and to realize data dimension reduction. The feature vector set was divided into different
 clusters by the K-means algorithm to achieve color classification and, thus, the solid wood panels with
 


 similar surface color were classified into one category. Furthermore, during twice clustering based
 
 on second-order color moment, texture recognition was realized on the basis of color classification. A
Citation: Wang, Z.; Zhuang, Z.; Liu, sample of beech wood was selected as the research object, not only was color classification completed,
Y.; Ding, F.; Tang, M. Color but texture recognition was also realized. The experimental results verified the effectiveness of the
Classification and Texture technical proposal.
Recognition System of Solid Wood
Panels. Forests 2021, 12, 1154. Keywords: solid wood board; color feature; unsupervised learning; color classification; texture
https://doi.org/10.3390/f12091154 recognition

Academic Editors: Michele Brunetti,
Michela Nocetti and
Alexander Petutschnigg
 1. Introduction
Received: 31 July 2021 Solid wood panels are widely used in solid wood furniture [1], wood flooring, and
Accepted: 23 August 2021 other industries because of their gloss finish, good decoration performance, effective sound
Published: 26 August 2021 absorption, high strength, easy processing, durability, and long service life. Moreover,
 waste wooden products can be degraded naturally and will hardly produce pollution. The
Publisher’s Note: MDPI stays neutral surface parameters of solid wood panels include color [2], texture [3], gloss [4], roughness,
with regard to jurisdictional claims in deformation rate, planeness, etc., which are directly related to the visual beauty and
published maps and institutional affil- decorative performance of wood products and are closely related to the quality evaluation
iations. of wood products. Therefore, it is theoretically and practically important to achieve
 automatic and intelligent detection and sorting of solid wood panel colors.
 Color is an important surface characteristic parameter of solid wood panels, as well
 as an important index to evaluate the quality, grade, and market value of wood products.
Copyright: © 2021 by the authors. In actual production, solid wood panels always need to be spliced together to form larger
Licensee MDPI, Basel, Switzerland. wood panels. It is preferable that panels with similar colors are spliced together to meet
This article is an open access article the needs of individual customers. Traditionally, the surface color classification of solid
distributed under the terms and wood has mainly been based on manual observation, which is significantly influenced by
conditions of the Creative Commons human factors and has low efficiency and, therefore, cannot meet society’s needs in terms of
Attribution (CC BY) license (https:// processing automation and intelligence and human–computer interaction. The emergence
creativecommons.org/licenses/by/ of new detection technologies that depend on machines overcomes the shortcomings of
4.0/).

Forests 2021, 12, 1154. https://doi.org/10.3390/f12091154 https://www.mdpi.com/journal/forests

Forests 2021, 12, 1154 2 of 18

 manual sorting. Lu [5] researched an automatic color sorting system for hardwood edge-
 glued panel parts, capable of sorting red oak panel parts into a number of color classes at
 plant production speeds and the test results showed that the qualified rate exceeded 91%.
 Schmitt [6] improved the existing system based on fuzzy language rules and constructed a
 fuzzy language rule classifier for wood color classification. The results obtained with the
 new method showed a real improvement in the recognition rate compared to a Bayesian
 classifier. In a study by Mária [7], a new wood color measurement method was verified
 using digital images in CIE L* a* b* color space birch color measurement and analysis.
 Syafinaz [8] studied the relationship between wood color and formaldehyde emissions from
 plywood of seven tropical hardwood species. Anton [9] determined the chemical properties
 of logs and wood samples after boiling, analyzed the wood, and isolated holocellulose
 and Sievetz cellulose by attenuated total reflection Fourier transform infrared spectroscopy
 (ATR-FTIR). It was found that the qualitative and quantitative changes in hemicellulose
 extracts during the cooking of birch were closely related to the measured pH value and
 wood color.
 In recent years, machine vision technology [10,11] and machine learning algorithms [12,13]
 have become popular solutions for signal processing and have been applied to color image
 classification. At present, the main direction of wood color classification is to select the
 appropriate color features and to construct accurate classifiers [14]. Rong-Hui [15] extracted
 eight color features of images in HSV (hue, saturation, and value), HSL (hue, saturation,
 and luminance), and HSI (hue, saturation, and intensity) color spaces and used the k-
 NN algorithm to realize the classification of farmland images in different environments.
 Xing [16] proposed a wood color classification method based on wood image features
 and a support vector machine (SVM). The hue and color vector angle (CVA) of the Bessel
 color system were used to characterize the wood color of the sample, which could quickly
 and accurately estimate the wood color grade. Vahid [17] evaluated and compared the
 performance of an artificial neural network, i.e., SVM, and the Naive Bayes (NB) classifier
 in thermally treated wood classifications. Using color brightness parameters as the unique
 feature, the accuracies of the SVM and NB models were 0.960 and 0.949, respectively. The
 application of unsupervised learning [18,19] technology in wood detection is helpful to
 improve the quality and efficiency of wood processing. Lin [20] applied the k-means
 algorithm to the surface color clustering of solid wood panels, and used the clustering
 results for online classification, the experimental results verified the feasibility of the
 proposed mechanism. In addition, K-means algorithm has been applied in many aspects,
 such as color image segmentation [21], data analysis [22], multi-objective programming [23],
 and so on.
 However, in actual production, due to long panel and high image resolution require-
 ments, computational complexity increases and the color difference is small; therefore, it is
 difficult to find general classification rules. In view of these difficulties, in this study, we
 divided the solid wood panel into blocks, and since the surface information of small area
 panels is relatively uniform, it is conducive to classification. Due to the large and complex
 surface information of solid wood panels, RGB (red, green, blue), HSV (hue, saturation,
 value) and Lab (laboratory) color spaces commonly used in the field of machine vision
 technology were selected to extract feature vectors based on first-order color moments,
 second-order color moments, and color histogram peaks. Mohseni [24] selected a color
 histogram with three channels in the RGB and HSV color space as a single feature to
 identify induced emotions and studied the influence of picture vision on human emotions.
 In this study, to describe the color characteristics of solid wood panels, first-order and
 second-order color moments, as well as a color histogram peak, were selected and extracted
 from the R, G, B, L, a, b, H, S, and V channels of a solid wood panel image, resulting in
 27 extracted feature vectors that were used for the clustering study of solid wood panels.
 This reduced the data dimension and ensured that enough information was extracted
 for clustering. The K-means unsupervised learning method was used to realize the color
 clustering of solid wood panels. The K-means clustering algorithm is usually used when

Forests 2021, 12, 1154 3 of 18

 we have unlabeled data (without defined categories) and it clusters the given data into
 K-clusters based on the K-centroids [25]. Then, the influence of different color characteris-
 tics on the clustering results was analyzed to find the best feature combination. Through
 theoretical analysis and experiments to find the optimal K value, a twice clustering method
 was proposed to identify texture information and to optimize the clustering results.
 The overall objective of this study was to realize color classification and texture
 recognition of solid wood panels. In this paper, beech wood panels were selected as
 the research object, and the image characteristics of beech wood panels were analyzed.
 Based on the analysis, we proposed a new technical proposal for the color classification of
 wood panels. The framework included image acquisition, image preprocessing, feature
 extraction, unsupervised clustering, color classification, and texture recognition. The image
 preprocessing algorithm was used to subtract the superfluous background of the images of
 solid wood panels. Then, feature vectors were extracted from the processed images based
 on the first-order color moments, second-order color moments, and color histogram peaks.
 In the research of clustering, the feature vector sets were partitioned into different clusters
 by the K-means algorithm to realize color classification. Finally, texture recognition was
 realized based on color classification. Using machine vision technology [26] and digital
 image processing technology [27], in this study, we took a sample of beech wood as the
 research object and conduct clustering analysis on the color characteristics of the wood
 panel surface.
 The rest of this paper is organized as follows. Section 2 describes the materials and
 methods used in the sorting of solid wood panels. Section 3 analyzes and shows the
 experimental results of color classification and texture recognition. Section 4 discusses the
 advantages of this technical proposal in color classification and texture recognition of solid
 wood panels. Finally, the conclusion is given in Section 5.

 2. Materials and Methods
 2.1. Panels
 A sample of beech wood was selected as the research object for the surface color
 cluster analysis of solid wood panels. The experiments were carried out on beech from
 Europe, which can adapt to a wide range of climate and temperature. The wood panels
 needed for the experiment were purchased from the wood merchants in Nanjing, Jiangsu
 Province and can be bought in the market. The density of the beech was 0.70 g/cm3 and
 the moisture content was 10.2%. After chord cutting, longitudinal cutting, planning, and
 polishing, a 1000 mm × 100 mm × 10 mm (length × width × height) smooth surface panel
 was obtained.

 2.2. Imaging
 In actual production, color classification is performed on the image of a solid wood
 floor panel through several processing steps. The structure of the image acquisition system
 built in this study is shown in Figure 1. There were four main components to the acquisition
 system, namely the transmission device, a CCD camera, a camera fixing device, and a
 control device.
 The core device of the CCD camera in the image acquisition system was a Linea
 LA-GC-02K05B color line scan camera (Teledyne DALSA Co., Waterloo, ON, Canada). The
 specific parameters of this device are shown in Table 1. The camera adopts high-sensitivity
 CMOS (complementary metal oxide semiconductor) technology; the line frequency can
 reach up to 26 KHz, with high acquisition speed, low noise, high single-line resolution, and
 high sensitivity. To improve the imaging quality of the acquired image, it was necessary to
 reduce the influence of light conditions on the acquisition effect during the data acquisition
 process, thus ensuring uniform illumination and minimizing the light reflection in the data
 acquisition area. Therefore, a linear light source LCOL-304-w (HZN, shanghai, China) was
 selected to meet the requirement of uniform illumination in the single-line area required
 by the linear CCD camera. The transmission belt in the image acquisition system consists

Forests 2021, 12, 1154 4 of 18

 of two parts rotating at a uniform speed. They dragged the wood panels into the data
 acquisition area, and then sent the solid wood panels completing the image acquisition
 to the storage area. The control and sensing device included sensors and a control panel.
 The wood transmission speed and the start and stop of data acquisition were controlled
021, 12, x FOR PEER REVIEW through the control panel. When collecting data, the scanning frequency was set to 1280 4 o
 lines for processing once, and after collecting data on the front of each sheet, a 2048 ×
 18,000 pixel image of the solid wood panel with a depth of 8 bit was obtained.

 Figure 1.Figure
 Image1. Image acquisition
 acquisition device:(1)
 device: (1)transmission
 transmission device, (2) transmission
 device, belt, (3) control
 (2) transmission panel,
 belt, (3) control pa
 (4) CCD camera holder, (5) CCD camera, (6) sample solid wood floor panel, (7) photoelectric sensor,
 (4) CCD camera holder, (5) CCD camera, (6) sample solid wood floor panel, (7) photoelectric sen
 and (8) pressure wheel.
 and (8) pressure wheel.
 Table 1. Specific parameters of the CCD camera.

 The core device
 Parameter Name of the CCD camera in the image
 Parameter Parameter acquisition
 Name system
 Parameterwas a Linea L
 GC‐02K05B color Modelline scanLinea camera (Teledyne DALSA
 LA-GC-02K05B Spectrum Co., Waterloo,Color ON, Canada).
 ◦C
 specific parameters of this device are shown in Table 1. The camera adopts
 Sensor Technology CMOS Operating temperature 0–65 high‐sensi
 Resolution 2048 × 2 Pixel depth 8 bit
 ity CMOS (complementary
 Pixel Size metal
 7.04 ×oxide
 7.04 µmsemiconductor) technology;Upthe
 Horizontal frequency line
 to 26 frequency
 KHz
 reach up to 26 KHz, with high acquisition speed, low noise, high single‐line resolut
 and high2.3.sensitivity.
 PreprocessingTo improve the imaging quality of the acquired image, it was ne
 Images
 sary to reduce the influence
 The three-channel ofimage
 color lightofconditions
 the BMP formaton collected
 the acquisition
 by the imageeffect during the d
 acquisition
 system was converted into a gray image, and morphological
 acquisition process, thus ensuring uniform illumination and minimizing the operations such as erosion and
 light ref
 dilation were carried out. The canny edge operator was applied to detect the boundary of
 tion in the
 thepanel.
 dataThe acquisition area.
 closed operation usedTherefore,
 a convolutiona kernel
 linear of 7light source
 × 7, the numberLCOL‐304‐w
 of iterations (HZ
 shanghai,
 wasChina) was
 1, and then selected
 used adaptivetobinarization.
 meet the requirement of uniform
 Finally, the boundary of the woodillumination
 in the image in the
 was found and the panel image was cut out. Canny edge
 gle‐line area required by the linear CCD camera. The transmission belt in the detection [28] technology is image
 applied in image processing [29] and has been widely used in traffic accident detection [30],
 quisition system consists of two parts rotating at a uniform speed. They dragged the w
 crack automatic segmentation [31], intelligent traffic management [32], and other fields.
 panels into the data
 The solid wood acquisition
 panel image was area, and then
 extracted sent
 from the thebackground,
 black solid wood and panels completing
 the influence
 image acquisition
 of the background to thewasstorage
 removed.area. The control
 One example and is
 of the results sensing
 shown indevice
 Figure 2.included sens
 and a control panel. The wood transmission speed and the start and stop of data acqu
 tion were controlled through the control panel. When collecting data, the scanning
 quency was set to 1280 lines for processing once, and after collecting data on the fron
 each sheet, a 2048 × 18,000 pixel image of the solid wood panel with a depth of 8 bit w
 obtained.

 Table 1. Specific parameters of the CCD camera.

 Parameter Name Parameter Parameter Name Paramete

OR PEER REVIEW 5 of 18

 other
 Forests 2021, fields.
 The solid wood panel image was extracted from the black background, and5 of 18
 12, 1154

 the influence of the background was removed. One example of the results is shown in
 Figure 2.

 (a) (b)
 Figure 2. Example ofFigure
 the result: (a) original
 2. Example image;
 of the result: (b) image
 (a) original with
 image; the background
 (b) image removed.
 with the background removed.

 2.4. Color Features of Solid Wood Panel Images
 2.4. Color Features of Solid
 TheWood
 generalPanel Images
 descriptive methods of measuring color features include color histograms,
 color moments,
 The general descriptive color sets,ofcolor
 methods aggregation
 measuring vectors,
 color color correlation
 features graphs,histo‐
 include color etc. Since
 color distribution information is mainly concentrated in low-order moments, only the
 grams, color moments, color sets, color aggregation vectors, color correlation graphs, etc.
 first-order moments (mean), second-order moments (variance), and third-order moments
 Since color distribution information
 (skewness) of colors areissufficient
 mainlytoconcentrated
 express the color indistribution
 low‐order in moments,
 the image. Inonly
 addition,
 the first‐order moments (mean),
 color sets second‐order
 are an approximation moments
 of the (variance),
 color histograms and aggregation
 and color third‐ordervectors
 mo‐ are
 ments (skewness) ofan colors areofsufficient
 evolution to expresstherefore,
 the color histograms; the color distribution
 color in the image.
 histograms describe In
 color features
 that are more extensive and representative. The first-order
 addition, color sets are an approximation of the color histograms and color aggregation color moment µi and the
 second-order color moment σi are mathematically defined as follows:
 vectors are an evolution of the color histograms; therefore, color histograms describe color
 features that are more extensive and representative. The 1 first‐order
 N color moment and
 the second‐order color moment are mathematicallyNdefined
 µi = ∑ pi,j
 as follows:
 (1)
 j =1

 1 !1
 1, N 
2 2 (1) (2)
 N ∑ i,j
 σi = p − µi
 j =1

 where pi,j is the pixel value at the image coordinate, and its actual meaning represents the
 pixel mean and variance of a picture.
 Color sorting is mainly 1 used to ensure a consistent appearance of the solid(2) wood
 , 
 
 panel to meet consumer psychology needs. The RGB color space is the most basic and
 most commonly applied color space in computer digital image processing. The hue and
 saturation in the HSV color space are directly related to humans’ perceptions of color.
 where , is the pixel value
 The hue at the image
 and saturation coordinate,
 components directly and its actual
 correspond meaning
 to humans’ represents
 perceptions and their
 the pixel mean andpsychological
 variance ofresponses
 a picture. to color and, therefore, are widely used in color classification.
 Color sortingLab color space
 is mainly can be
 used to directly
 ensureused to compare appearance
 a consistent and analyze different
 of thecolors
 solidbywood
 using the
 geometric distance in the color space, which can be effectively
 panel to meet consumer psychology needs. The RGB color space is the most basic and and conveniently used to
 measure the smaller color difference. The distribution range of the wood color is narrow.
 most commonly applied color space in computer digital image processing. The hue and
 The color characteristics in the Lab color space are used to represent the surface color of
 saturation in the HSV color
 wood, whichspace are directly
 is beneficial relatedand
 for comparing to humans’
 classifyingperceptions
 wood surface of color. The
 colors.
 hue and saturation components
 The calculationdirectly
 formulascorrespond to humans’
 of the first-order color momentperceptions and theircolor
 and the second-order
 moment are shown in Equations (1) and (2). The average
 psychological responses to color and, therefore, are widely used in color classification.value of wood image pixels
 Lab color space can be directly used to compare and analyze different colors by using the
 geometric distance in the color space, which can be effectively and conveniently used to
 measure the smaller color difference. The distribution range of the wood color is narrow.

Forests 2021, 12, x FOR PEER REVIEW 6 of 18

 Forests 2021, 12, 1154 6 of 18
 The calculation formulas of the first-order color moment and the second-order color
 moment are shown in Equations (1) and (2). The average value of wood image pixels de-
 scribes thethe
 describes overall
 overallcolor
 color distribution
 distributionofofwood
 woodimages
 imagesandandthe
 thevariance
 variancedescribes
 describes the
 the uni-
 formity
 uniformity of the image distribution in the color domain. The color histogram of threecolor
 of the image distribution in the color domain. The color histogram of three
 spaces of solid
 color spaces woodwood
 of solid panelpanel
 image is shown
 image is shownin Figure 3, which
 in Figure shows
 3, which thethe
 shows numbers
 numbersof of
 pixels
 with
 pixelseach
 withvalue in theinR,the
 each value G, R,
 B, G,
 H,B,S,H,
 V,S,L,V,a,L,and b channels,
 a, and respectively.
 b channels, respectively.

 (a) Color histogram of RGB color space (b) Color histogram of HSV color space

 (c) Color histogram of Lab color space

 Figure 3.
 3. Color
 Colorhistogram
 histogramofofbeech
 beechpanel
 panelinin RGB,
 RGB, HSV,
 HSV, andand
 LabLab color
 color space.
 space.

 2.5. Solid Wood Panel Clustering
 2.5. Solid Wood Panel Clustering
 Traditional artificial methods are not very effective for color classification of solid
 Traditional artificial methods are not very effective for color classification of solid
 wood panel surface colors, while supervised learning consumes a lot of manpower and
 wood panel surface colors, while supervised learning consumes a lot of manpower and
 material resources for label setting. In this study, an unsupervised learning clustering
 material
 algorithmresources for label
 was introduced setting.
 to solve theIn thisclassification
 color study, an unsupervised learning
 problem of solid woodclustering
 panels. al-
 gorithm was introduced to solve the color classification problem of solid wood panels.
 2.5.1. Dataset
 2.5.1.The
 Dataset
 images of solid wood panels collected by the image acquisition system had high
 resolutions and were
 The images large
 of solid in size.
 wood The collected
 panels original image
 by thesize
 imageof each collected
 acquisition imagehad
 system washigh
 2048 * 18,000 pixels and contained a variety of information such as color,
 resolutions and were large in size. The original image size of each collected image texture, and was
 defects in the image. As shown in Figure 4, a variety of complex information co-exists
 2048 * 18,000 pixels and contained a variety of information such as color, texture, and in an de-
 image,
 fects inwhich makesAs
 the image. color classification
 shown in Figuredifficult. Therefore,
 4, a variety in thisinformation
 of complex study, the larger imagein an
 co-exists
 was divided into blocks, and each large image was divided into 3 × 3 = 9 subgraphs.
 image, which makes color classification difficult. Therefore, in this study, the larger Due
 image
 to the small area of each subgraph, the surface color information was more centralized
 was divided into blocks, and each large image was divided into 3 × 3 = 9 subgraphs. Due
 and unified. Such a dataset was conducive to the classification of the solid wood panel
 to the small area of each subgraph, the surface color information was more centralized
 surface colors.
 and unified. Such a dataset was conducive to the classification of the solid wood panel
 A sample of beech wood was selected as the research object for the surface color
 surface colors.
 cluster analysis of solid wood. A total of 200 samples of beech wood were preprocessed
 and segmented to obtain the dataset of 200 × 3 × 3 = 1800.

Forests
 Forests2021,
 2021,12,
 12,x1154
 FOR PEER REVIEW 7 7ofof18
 18

 Figure 4. Original image and divided graph.
 Figure 4. Original image and divided graph.
 2.5.2. Clustering Based on Color Features
 A sample of beech wood was selected as the research object for the surface color clus‐
 Firstly, as previously mentioned, the solid wood panel images were preprocessed.
 ter analysis of solid wood. A total of 200 samples of beech wood were preprocessed and
 Then, the first-order color moments, second-order color moments, and color histogram
 segmented to obtain the dataset of 200 × 3 × 3 = 1800.
 peaks of the R, G, B, L, a, b, H, S, and V nine color channels of each solid wood panel
 image were extracted, and a total of 27 color features were used for clustering. Finally,
 2.5.2. Clustering Based on Color Features
 the k-means clustering algorithm was selected to realize the clustering of surface color
 Firstly, as previously
 characteristics of solid wood. mentioned, the solid wood panel images were preprocessed.
 Then,Thethe steps
 first‐order
 of the color moments,
 algorithm are tosecond‐order
 divide the data color
 intomoments,
 K groups,and color histogram
 to randomly select K
 peaks
 objects as the initial clustering center, then to calculate the distance between each panel
 of the R, G, B, L, a, b, H, S, and V nine color channels of each solid wood object
 image were
 and each extracted,
 seed clusteringandcenter,
 a total and
 of 27tocolor features
 assign were used
 each object to thefor clustering.
 nearest Finally,
 clustering the
 center.
 k‐means clustering algorithm was selected to realize the clustering
 Clustering centers and the objects assigned to them represent a clustering. The cluster of surface color char‐
 acteristics of solid wood.
 center is recalculated based on the existing objects in the cluster when each sample is
 The steps
 assigned. Thisofprocess
 the algorithm
 is repeatedare tountil
 divide the dataare
 no objects into K groups,to
 reassigned todifferent
 randomly select K
 clustering
 objects
 centersasorthe
 wheninitial
 the clustering
 clustering center,
 centers then
 do notto change.
 calculate the distance between each object
 and each
 The seed clustering
 K-means center,
 clustering and toisassign
 method widelyeachusedobject
 and istostable
 the nearest clustering
 and effective. center.
 However,
 Clustering centers and the objects assigned to them represent a clustering.
 the algorithm needs to specify the K value in advance. Since the clustering is unsupervised, The cluster
 center is recalculated
 the optimal based on
 K value cannot be the existing objects
 determined based on in the
 the classification
 cluster when results.
 each sample
 Thereishave
 as‐
 signed. This process
 been studies on the is repeated
 optimal until no
 K value objects
 [33,34]. In are
 thisreassigned to different
 study, the contour clustering
 coefficient cen‐
 method
 wasorselected
 ters when the to determine the K value
 clustering centers do not ofchange.
 the clustering of the solid wood panel images.
 The The
 coreK‐means
 index of this methodmethod
 clustering is the silhouette
 is widely coefficient
 used and is[35]. Theand
 stable silhouette coefficient
 effective. However, of
 a sample
 the algorithm Xi is defined
 pointneeds to specify as the
 follows:
 K value in advance. Since the clustering is unsuper‐
 vised, the optimal K value cannot be determined based on the classification results. There
 have been studies on the optimal K value b−a
 S = [33,34]. In this study, the contour coefficient (3)
 method was selected to determine the K value( a, max ofbthe
 ) clustering of the solid wood panel
 images.
 where aThe core
 is the index distance
 average of this methodbetween is the silhouette
 Xi and coefficient
 other samples in [35]. The silhouette
 the same co‐
 cluster, called
 efficient
 cohesion, of and
 a sample
 wherepoint Xi isaverage
 b is the defineddistance
 as follows:between X and all samples in the nearest
 i
 cluster is as follows:
 cluster, called separation. The definition of nearest
 S (3)
 max , 
 1
 n P∑
 2
 j = arg min
 where a is the average distanceCbetween Xi | P −samples
 Xi | (4)
 Ck and other
 ∈Ck
 in the same cluster, called
 cohesion, and where b is the average distance between Xi and all samples in the nearest
 cluster,
 where Pcalled
 is theseparation.
 sample in aThe definition
 cluster Ck . of nearest cluster is as follows:
 In fact, after taking the average distance 1 of Xi to all samples of a certain cluster
 as the distance from the point to thearg min one | 
 cluster, | closest to Xi is selected as (4)
 cluster the
 
 nearest cluster. ∈

 whereThe
 P isaverage contour
 the sample in a coefficient
 cluster Ck. is obtained by calculating the contour coefficient of all
 samples and then by calculating the average value. The closer the distance between the

In fact, after taking the average distance of Xi to all samples of a certain cluster as the
 distance from the point to the cluster, one cluster closest to Xi is selected as the nearest
 cluster.
Forests 2021, 12, 1154 The average contour coefficient is obtained by calculating the contour coefficient 8 of 18 of
 all samples and then by calculating the average value. The closer the distance between the
 samples in one cluster, the farther the distance between the samples in different clusters,
 the greater
 samples the cluster,
 in one averagethecontour coefficient,
 farther the and thethe
 distance between better
 samplesthein
 clustering result. The
 different clusters, the maxi‐
 greater
 mum the average
 average contour
 contour coefficient,
 coefficient and
 K is thethebest
 better the clustering
 clustering result.
 number. The maximum
 Using the dataset pro‐
 average
 duced in contour coefficient
 this scheme K is the bestthat
 and considering clustering number.
 the K value canUsing
 rangethe dataset
 from 2 to produced
 10, each K value
 was clustered and the corresponding contour coefficient was obtained. K
 in this scheme and considering that the K value can range from 2 to 10, each valuethe
 Then, wasrelation‐
 clustered and the corresponding contour coefficient was obtained. Then, the relationship
 ship between the K value and contour coefficient was plotted, and the K value with the
 between the K value and contour coefficient was plotted, and the K value with the largest
 largest
 contourcontour coefficient
 coefficient was The
 was selected. selected.
 pythonThe python implementation
 implementation is shown
 is shown in Figure 5, andinthe
 Figure 5,
 and the optimal
 optimal K value
 K value was 8. was 8.

 Figure 5. Contour coefficient and K value.
 Figure 5. Contour coefficient and K value.
 The best K value obtained by theoretical analysis was 8, but in actual production, it is
 oftenThe
 adjusted
 best Kaccording to different
 value obtained customer needs.
 by theoretical In order
 analysis was to 8, meet theactual
 but in actualproduction,
 needs, it
 the K value in the experiment was set to 3–11, and the dataset of solid wood was clustered
 is often adjusted according to different customer needs. In order to meet the actual needs,
 with
 the K different
 value in color feature combinations.
 the experiment was set to 3–11, and the dataset of solid wood was clustered
 In this study, three color features were selected and combined as “first-order color
 with different color feature combinations.
 moment’, “first-order color moment + second-order color moment’, “first-order color
 In this study, three color features were selected and combined as “first‐order color
 moment + color histogram peak”, “second-order color moment + color histogram peak”,
 moment’, “first‐order
 and “first-order color moment
 color moment + second‐order
 + second-order color+moment’,
 color moment “first‐order
 color histogram peak”. color mo‐
 ment + color histogram peak”, “second‐order color moment + color histogram peak”, and
 3. Results color moment + second‐order color moment + color histogram peak”.
 “first‐order
 3.1. Color Classification of Solid Wood Panels
 3. Results
 Figure 6 shows that, under the five color feature combinations, the solid wood panels
 in the dataset were clustered into three, four, five, six, and seven categories. The number of
 3.1. Color Classification of Solid Wood Panels
 columns in each figure represents the number of clusters (K value), and five rows represent
 five Figure 6 shows
 color feature that, under
 combinations. Asthe five color
 a single imagefeature
 was not combinations, the
 representative, in solid
 order to wood
 reflect panels
 inthe
 the dataset
 real were clustered
 classification into
 effect of the three, four,clustering
 unsupervised five, six, algorithm
 and sevenoncategories.
 the surfaceThecolornumber
 ofofcolumns
 the solid in
 woodeach figure
 panel, ninerepresents
 images weretherandomly
 number of clusters
 selected from(Keach
 value), and five
 clustering rows rep‐
 result
 and spliced together to form a new image to represent the overall performance
 resent five color feature combinations. As a single image was not representative, in order of this class.
 toUnder thethe
 reflect combination of five color
 real classification features,
 effect of thefor each K value, clustering
 unsupervised there were five × K images.
 algorithm on the sur‐
 Finally, images under each color feature combination were arranged in order from light
 face color of the solid wood panel, nine images were randomly selected from each clus‐
 to dark.
 tering result and spliced together to form a new image to represent the overall perfor‐
 mance of this class. Under the combination of five color features, for each K value, there
 were five × K images. Finally, images under each color feature combination were arranged
 in order from light to dark.

Forests 2021, 12, 1154 9 of 18
Forests 2021, 12, x FOR PEER REVIEW 9 of 18

 (a) K = 3 (b) K = 4

 (c) K = 5 (d) K = 6

 (e) K = 7
 Figure
 Figure6.6.KK= =3–7
 3–7clustering
 clusteringresult
 resultimages.
 images.

Forests 2021, 12, x FOR PEER REVIEW 10 of 18

 Forests 2021, 12, 1154 10 of 18

 It can be seen from each row of the above figures that, under a certain color feature
 combination, although the number of clustering the centers, namely the K value, were
 It can be seen from each row of the above figures that, under a certain color feature
 different, the clustering results were able to sort solid wood panel surface colors from light
 combination, although the number of clustering the centers, namely the K value, were
 to dark on the whole. With an increase in K value, the difference in surface color infor‐
 different, the clustering results were able to sort solid wood panel surface colors from
 mation of dark
 light to the solid
 on thewood
 whole. panel
 Withinaneach classingradually
 increase K value, the decreases.
 differenceFrom each column,
 in surface color it
 caninformation
 be seen that thesolid
 of the surface
 woodcolor
 panelinformation
 in each class of solid wood
 gradually panels
 decreases. in each
 From class was
 each column, it cha‐
 otic, that some contained obvious texture information, and that some classified
 can be seen that the surface color information of solid wood panels in each class was chaotic, the panels
 with
 thatdifferent depthsobvious
 some contained into one class information,
 texture (the last column
 and thatof some
 each classified
 graph isthe thepanels
 most with
 obvious).
 Thedifferent
 key to depths
 the aboveinto problems
 one class (the last column
 is that, due tooftheeach graph
 low is the setting,
 K value most obvious). The keyinitial
 the different
 to the above
 clustering problems
 centers mayislead
 that, to
 duethe
 to the low K value
 instability setting,
 of the the different
 clustering initial
 results, clustering
 which produces
 centers may lead to the instability of the clustering results, which produces multiple local
 multiple local optimal values.
 optimal values.
 When the clustering center was selected as the theoretical optimal value, i.e., K = 8, it
 When the clustering center was selected as the theoretical optimal value, i.e., K = 8, it
 cancan
 bebeseen
 seenfrom
 fromFigure
 Figure 77that
 thatthe
 theclustering
 clustering results
 results improved
 improved compared
 compared with thewith the previ‐
 previous
 ousfive
 five
 KK values.
 values. As Asthe the number
 number of clusters
 of clusters increased
 increased as the Kasvalue
 the Kincreased,
 value increased, the simi‐
 the similarity
 larity of color information on the surface of solid wood panel in each
 of color information on the surface of solid wood panel in each class continually increased class continually
 increased
 and the and the difference
 difference of color information
 of color information on the surfaceon ofthesolid
 surface
 woodofpanel
 solidinwood panel in
 adjacent
 classes gradually decreased.
 adjacent classes gradually decreased.

 Figure
 Figure7.7.KK==88clustering resultimages.
 clustering result images.

 Table
 Table 2 2shows
 showsthe
 thenumber
 number ofof solid
 solidwood
 woodpanel
 panelimages
 imagesin each cluster
 in each under
 cluster five color
 under five color
 feature combinations. It can be seen that, for the dataset of 1800 images, when K = 8,
 feature combinations. It can be seen that, for the dataset of 1800 images, when K = 8, therethere
 were only 11 images in the least number of classes and that their surface color information
 were only 11 images in the least number of classes and that their surface color information
 was very similar.
 was very similar.

 Table 2. K = 8 number of images in each cluster.

 Features 1 2 3 4 5 6 7 8
 Mean 362 330 328 251 239 152 109 29
 Mean + variance 393 388 319 247 231 111 100 11
 Mean + histogram 381 348 307 272 207 160 90 35
 Variance + histogram 377 366 313 289 191 165 87 12

Forests 2021, 12, 1154 11 of 18

 Table 2. K = 8 number of images in each cluster.

 Features 1 2 3 4 5 6 7 8
 Mean 362 330 328 251 239 152 109 29
 Mean + variance 393 388 319 247 231 111 100 11
 Mean + histogram 381 348 307 272 207 160 90 35
 Variance + histogram 377 366 313 289 191 165 87 12
 Mean + variance + histogram 384 358 311 266 190 163 82 46

 The theoretical analysis has certain guiding significance, but in actual production, we
 should make appropriate adjustments according to the complicated information of the
 surface color of solid wood panel. Previous studies and analyses have found that, when
 the K value is less than the theoretical optimal, the clustering results only realize the rough
 classification of the solid wood panel surface color and the difference in the same type of
 panel is large, which cannot meet the actual needs. When the theoretical optimal K value is
 selected, the classification effect is improved and the difference of the same type of panel
 decreases. In order to find the K value that is most suitable for solid wood panel surface
 color classification, it is necessary to add a K value for further research.
 It has been found that the similarity of the same kind of panel increases with an
 increase in K value, which is a positive correlation. In order to find the number of clustering
 centers most suitable for solid wood panels, we continued to increase the K values to 9, 10,
 and 11 and analyzed their classification effect.
 The experimental results of K = 9, 10, and 11 showed the following: Firstly, the color of
 solid wood panels in each class could be sorted from light to dark, especially based on the
 first-order color moment or color histogram peak clustering, and the results were smoother
 and less hierarchical. Secondly, as the K value increased, the similarity of images in each
 class increased and the difference between classes decreased. Thirdly, Lines 2, 4, and 5 of
 Figures 8–10 appeared to be obvious classes of image texture information. Fourthly, from
Forests 2021, 12, x FOR PEER REVIEWTables 3–5, the number in the last column changed a little, and the number in the previous
 12 of 18
 columns changed slightly.

 Figure 8. K == 99 clustering
 clustering result
 result images.

Forests 2021, 12, 1154 12 of 18

Forests 2021, 12, x FOR PEER REVIEW 12 of 18
 Table 3. K = 9 number of images in each cluster.

 Features 1 2 3 4 5 6 7 8 9
 Mean 325 312 304 242 239 182 109 77 10
 Mean + variance 384 325 305 253 221 110 103 88 11
 mean + histogram 371 336 308 232 197 178 90 79 9
 Variance + histogram 349 343 284 231 228 154 114 86 11
 Mean + variance + histogram 375 345 309 244 190 160 92 76 9

 Table 4. K = 10 number of images in each cluster.

 Features 1 2 3 4 5 6 7 8 9 10
 Mean 288 269 258 232 209 191 166 95 81 11
 Mean + variance 316 301 265 243 205 179 108 94 78 11
 Mean + histogram 310 299 295 229 221 137 154 77 69 9
 Variance + histogram 363 335 289 259 194 176 86 80 11 7
 Mean + variance + histogram 304 287 284 228 208 178 146 77 77 11

 Table 5. K = 11 number of images in each cluster.

 Features 1 2 3 4 5 6 7 8 9 10 11
 Mean 265 265 256 235 199 179 146 122 68 55 10
 Mean + variance 300 298 262 226 205 175 108 96 75 44 11
 Mean + histogram 274 270 257 214 209 170 149 125 77 46 9
 Variance + histogram 316 312 244 243 189 152 109 85 73 66 11
 Mean + variance + histogram 316 300 287 243 201 148 136 72 48 40 9
 Figure 8. K = 9 clustering result images.

 Figure
 Figure9.9.KK==10
 10 clustering resultimages.
 clustering result images.

 It can be seen from Figures 8–10 that different color feature combinations have dif-
 ferent effects on clustering. Color and texture are important characteristics that affect the
 appearance of wood products. Color classification and texture recognition are conducive
 to improving the effect of solid wood panel sorting.

Forests 2021, 12, 1154 13 of 18
Forests 2021, 12, x FOR PEER REVIEW 13 of 18

 Figure 10.
 Figure 10. K = 11 clustering result images.
 When K ≥ 9, there was almost no change in the clustering results of the last class of
 It can be seen from Figures 8–10 that different color feature combinations have dif‐
 images, that is, the clustering centers of these images did not change. With an increase in K
 ferent effects on clustering. Color and texture are important characteristics that affect the
 value, the similarity of the images was higher and higher. Therefore, the K value, in this
 appearance of wood products. Color classification and texture recognition are conducive
 study, should be set to nine.
 to improving the effect of solid wood panel sorting.
 When K
 3.2. Texture ≥ 9, thereRecognition
 Information was almost no change in the clustering results of the last class of
 images,
 Forthat is, the clustering
 the classes with obviouscenters of these
 texture images did
 information not change.
 in Lines With
 2, 4, and 5 inan increase
 Figures in
 8–10,
 K value, the similarity of the images was higher and higher. Therefore, the K value,
 we found that their color feature combinations had second-order color moments. In total, in this
 study,
 166 should
 images in be
 theset to nine.category of “mean” combination in Table 4 were selected as
 seventh
 datasets, and the second-order color moments were used to extract feature vectors for
 3.2. Textureclustering
 secondary Information
 ofRecognition
 these images, K = 2.
 For the classes
 In Figure with
 11, nine obvious selected
 randomly texture information
 images wereintaken
 Linesfrom
 2, 4, 166
 andimages
 5 in Figures
 of the8–10,
 first
 we found that their color feature combinations had second‐order color moments.
 clustering, where b and c are the results of the second clustering—in b, the texture was In total,
 166 images
 relatively in theand
 smooth seventh category
 the color of “mean”
 information combination
 was unified, while in Table
 in c, 4 wereinformation
 the texture selected as
 datasets,
 was and the Itsecond‐order
 prominent. can be seen color moments
 that the were used
 second-order to moment
 color extract feature
 had a vectors for
 significant
 secondary
 influence on clustering of these
 the texture images,
 recognition ofKsolid
 = 2. wood.
 In Figure 11, nine randomly selected images were taken from 166 images of the first
 clustering, where b and c are the results of the second clustering—in b, the texture was
 relatively smooth and the color information was unified, while in c, the texture infor‐
 mation was prominent. It can be seen that the second‐order color moment had a signifi‐
 cant influence on the texture recognition of solid wood.

 Figure 11. Second-order color moment result images: (a) the first clustering result images, (b) the
 second clustering result images with smooth outer surface, (c) the second clustering result images
 Figure
 with 11. Second‐order
 obvious color moment result images: (a) the first clustering result images, (b) the
 texture information.
 second clustering result images with smooth outer surface, (c) the second clustering result images
 with obvious texture
 The images information.
 of solid wood panels for the first eight categories of the “mean + histogram”
 combination in Table 3 (the last category had fewer images and was relatively fixed)
 wereThe images
 selected of solid
 as the wood
 datasets, panels
 and for thevectors
 the feature first eight categories
 extracted of second-order
 by the the “mean + histo‐
 color
 gram” combination
 moments in Table
 were used for 3 (theclustering
 the second last category haddatasets.
 of these fewer images and was relatively

Forests 2021, 12, x FOR PEER REVIEW 14 of 18

Forests 2021, 12, 1154 14 of 18

 fixed) were selected as the datasets, and the feature vectors extracted by the second‐order
 color moments were used for the second clustering of these datasets.
 The experimental results are are shown
 shown inin Figure
 Figure 12.
 12. The second clustering divided
 divided each
 each
 relatively smooth,
 result of the first clustering into two categories; one category was relatively smooth, with
 with
 uniform color information, and the other category had prominent texture information information and
 and
 complexsurface
 more complex surfaceinformation.
 information.After
 Afterthe
 the overall
 overall color
 color classification
 classification of the
 of the images
 images in
 in the
 the dataset
 dataset waswas completed,
 completed, the the second‐order
 second-order color
 color moment
 moment effectively
 effectively identified
 identified the the tex‐
 texture
 ture information
 information in each
 in each class class and further
 and further optimized
 optimized the clustering
 the clustering results.
 results.

 Figure 12. Two‐clustering result images.
 Two-clustering result images.

 When
 When the the clustering
 clustering center
 center K K== 99 was
 was selected,
 selected, the
 the clustering
 clustering results
 results of
 of the
 the surface
 surface
 information
 information of solid wood panels met the production requirements; therefore, the optimal
 of solid wood panels met the production requirements; therefore, the optimal
 K
 K value
 value in in this
 this experiment
 experiment was was nine. In actual
 nine. In actual production,
 production, ifif only
 only aa rough
 rough classification
 classification ofof
 surface
 surface color
 color of
 of solid
 solid wood
 wood panels
 panels isis needed,
 needed, aa smaller
 smaller KK value
 value can
 can be be selected.
 selected. InIn order
 order
 to
 to achieve
 achieve aa more
 more detailed
 detailed color
 color classification,
 classification, an
 an appropriate
 appropriate feature
 feature combination
 combination can can be
 be
 selected along with an increase of the K value of clustering center. If we
 selected along with an increase of the K value of clustering center. If we want to identifywant to identify
 the
 the texture
 texture information,
 information, we we can
 can choose
 choose “twice
 “twice clustering”
 clustering” and
 and further
 further refine
 refine the
 the results
 results of
 of
 the
 the first clustering. In practical applications, different processing methods can be selected
 first clustering. In practical applications, different processing methods can be selected
 to
 to meet
 meet thethe personalized
 personalized needs
 needs of
 of different
 different customers.
 customers.
 In
 In this experiment, the data in the last column
 this experiment, the data in the last column in in Tables
 Tables 3–5
 3–5 corresponded
 corresponded to to the
 the
 darkest color class
 darkest color class in
 in aagraph
 graphwith
 withthe
 thesame
 sameKKvalue.
 value.TheThetypes
 types and
 and numbers
 numbers of of images
 images in
 in these categories basically did not change with a change in color feature combination
 these categories basically did not change with a change in color feature combination and
 and K value. We found that these specific images were quite different from other images,
 K value. We found that these specific images were quite different from other images,
 which may have been caused by pests, diseases, or special growth environments and can
 which may have been caused by pests, diseases, or special growth environments and can
 be treated as exceptions.
 be treated as exceptions.
 From Figures 8–10, it can be seen that, after sorting the clustering results from light to
 From Figures 8–10, it can be seen that, after sorting the clustering results from light
 dark, the classification of solid wood surface color was realized. The larger the K value, the
 to dark, the classification of solid wood surface color was realized. The larger the K value,
 smaller the difference between grades and the more detailed the classification. It can be
 the smaller the difference between grades and the more detailed the classification. It can
 seen from Figure 12 that the texture information was recognized by using texture sensitive
 be seen from Figure 12 that the texture information was recognized by using texture sen‐
 color features. After color classification and texture recognition, the panels with similar
 sitive color features. After color classification and texture recognition, the panels with sim‐
 surface information were spliced together to realize the coordination and consistency of
 ilar surface information were spliced together to realize the coordination and consistency
 the appearance of the spliced panel, as shown in Figure 13.
 of the appearance of the spliced panel, as shown in Figure 13.

ests 2021, 12, x FOR
 Forests PEER
 2021, 12, REVIEW
 1154 15 of 18 15 of 18

 (a) No clustering

 (b) After clustering
 Figure 13. The result
 Figure 13.images: (a) no
 The result classification,
 images: (b) after clustering.
 (a) no classification, (b) after clustering.

 4. Discussion4. Discussion
 The above The aboveverified
 sections sectionsthe verified the effectiveness
 effectiveness of solidofwoodsolid wood
 surface surface
 color color
 sortingsorting based
 on machine vision and unsupervised learning algorithm.
 based on machine vision and unsupervised learning algorithm. By creating new datasets By creating new datasets and
 and analyzing the effect of different color feature combinations, a new classificationmethod for
 analyzing the effect of different color feature combinations, a new classification
 method formulti-level
 multi‐levelcolor colorsorting
 sortingof solid wood
 of solid woodplates was was
 plates proposed, and texture
 proposed, recognition was
 and texture
 recognition was realized. In this study, the problem of setting the K value in advancesolved
 realized. In this study, the problem of setting the K value in advance was was through
 theoretical analysis and experimental summary.
 solved through theoretical analysis and experimental summary.
 Based on learning style, machine learning can be divided into supervised learning
 Based on learning style, machine learning can be divided into supervised learning
 and unsupervised learning. At present, the supervised learning classification algorithms
 and unsupervised learning. At present, the supervised learning classification algorithms
 that are widely used mainly include decision tree, k-nearest neighbor (KNN) algorithm,
 that are widely used mainly include decision tree, k‐nearest neighbor (KNN) algorithm,
 support vector machine (SVM) algorithm, naive Bayesian Model (NBM), deep learning,
 support vector machine (SVM) algorithm, naive Bayesian Model (NBM), deep learning,
 and other methods. However, the surface information of solid wood panels is complex,
 and other methods. However, the surface information of solid wood panels is complex, it
 it is difficult to label categories manually, and the cost of manual category labeling is too
 is difficult to label categories manually, and the cost of manual category labeling is too
 high to obtain enough prior knowledge for supervised learning. Unsupervised learning
 high to obtain enough prior knowledge for supervised learning. Unsupervised learning
 technology can effectively solve this problem.
 technology can effectively solve this problem.
 In this paper, the feature vectors of solid wood panel images were extracted based on
 In this paper, the feature vectors of solid wood panel images were extracted based
 color features for clustering, which realized data dimensionality reduction, which can not
 on color features for clustering,
 only reduce which
 the difficulty ofrealized databut
 calculation, dimensionality reduction,
 also save processing which
 time [36].canIn the field of
 not only reduce the difficulty of calculation, but also save processing
 machine learning, there are also Hough transform method and wavelet and time [36]. In the fieldlocal binary
 of machinepattern
 learning, there are
 method foralso Hough
 feature transform method and wavelet and local binary
 extraction.
 pattern method Compared
 for feature with extraction.
 the supervised learning method, the unsupervised learning method
 Compared
 appliedwith to the
 the supervised learningofmethod,
 color classification solid wood the unsupervised
 panels can overcomelearning themethod
 disadvantages of
 applied to difficult
 the colormanual
 classification
 category of labeling
 solid wood andpanels
 high cost canofovercome the disadvantages
 manual category labeling, because the
 of difficult supervised
 manual category learning labeling and needs
 algorithm high costenoughof manual category labeling,
 prior knowledge becauselearning.
 for supervised
 the supervised learning algorithm needs enough prior knowledge
 Feature vectors were extracted based on the selected color features. On for supervised learn‐
 the one hand,
 ing. sufficient information of solid wood panels can be obtained; on the other hand, it can reduce
 Featurethevectors
 dimensionwere extracted based on
 of data, reduce thethe selectedofcolor
 difficulty features.and
 calculation, On the one hand,
 improve the processing
 sufficient information of solid according
 speed. In addition, wood panels to thecanclustering
 be obtained; on the
 results other hand,
 of different coloritfeatures,
 can re‐ we made
 duce the dimension
 it clear which of data,
 color reduce
 feature the
 wasdifficulty
 suitable for ofcolor
 calculation, and improve
 classification and which thecolor
 pro‐feature was
 cessing speed. In addition,
 suitable for texture according to the clustering results of different color features,
 recognition.
 we made it clearInwhich termscolor feature was results,
 of experimental suitablefrom for color classificationeffect,
 the classification and which color scheme
 the technical
 feature wasproposed
 suitable for texture recognition.
 in this paper can realize more detailed classification with the increase of K value.
 In terms
 Thisofisexperimental results, by
 difficult to complete from the classification
 supervised effect, the
 learning method. Duetechnical scheme
 to the fact that the surface
 proposed ininformation
 this paper can realize
 of solid morepanels
 wood detailed wasclassification
 complex and with the increase
 there of K value.
 were similarities between the
 This is difficult to complete
 panels, with the by supervised
 increase of thelearning
 number method. Due to the
 of classifications, thefact that the surface
 difference between different
 informationclasses
 of solid wood panels
 decreased. It is was complex
 difficult and theresuch
 to distinguish weresmall
 similarities
 differencesbetweenin thetheprogress of
 panels, with manual category
 the increase labeling,
 of the number andofmanual labelingthe
 classifications, cannot provide
 difference a training
 between set, test set, and
 different
 verification
 classes decreased. It isset for theto
 difficult supervised
 distinguish learning
 such smallalgorithm.
 differences in the progress of