RU2444002C2 - Method and device to control saw timber - Google Patents

Abstract

FIELD: instrument making.

SUBSTANCE: image generation facility (8) captures a colour image of saw timber (9). An image processing facility (1) receives a colour distribution of a colour image captured by the image generation facility (8), compares histograms of brightness of the produced colour distribution with the predetermined histogram of brightness of the colour distribution of normal saw timber. It evaluates the received colour distribution as abnormal, when it deviates from the colour distribution of normal saw timber by a preset value or more, and detects a defect of saw timber having abnormal colour distribution deviating by the value that exceeds the preset value.

EFFECT: improved accuracy of evaluation.

7 cl, 10 dwg

Description

The present invention relates to a method for controlling lumber and a device for its implementation, designed to detect defective areas due to color changes that exist in wood-based materials, such as veneers or blanks cut from a wooden log or the like. For example, in the manufacture of glued plywood, a log is cut using a cutting tool to sequentially produce veneer sheets several millimeters thick. Then the veneer sheets are cut to a predetermined size and, after drying, several veneer sheets are fastened together by lamination using glue. During these technological processes, it is necessary to sort the veneer sheets into sheets that will go to the outer layers of glued plywood, or, in other words, sheets with small or aesthetically insignificant defects and sheets that will go to the inner layers, or, in other words, sheets with many, although not problematic, but aesthetically significant defects, in accordance with the location, quantity, size, etc. defects that affect the quality of the veneer (for example, defects formed due to a change in the color of the surface of the veneer, deformation, holes in the veneer, formed by fallen knots, as well as cracks). Such sorting includes classification, for example, into five to seven varieties.

State of the art

Typically, the process of sorting veneers into sheets that will go to the surface layers of glued plywood, and sheets that go to its inner layers, involves determining with the naked eye the quality of the veneer sheets transported by the conveyor.

In addition, a known method for monitoring defective areas in lumber includes: capturing an image of lumber using a color camera on charge-coupled devices (CCD); comparing the video signal with the reference colors of the resin and the reference samples of color changes using the device for extracting a color image; and converting the video signal to binary, marking the binary converted image corresponding to the area being monitored, and matching the binary converted image with a valid value to detect defective portions, including, for example, tar, rot and color changes (see Patent Document 1 )

Patent Document 1: Japanese Patent Application Laid-Open No. 09-210785.

SUMMARY OF THE INVENTION

Problems Solved by the Invention

The following problems are associated with the known methods described above.

Definition of quality with the naked eye is inaccurate due to the scatter in assessments by different people and is characterized by low productivity, since it is impossible to increase the speed of the conveyor.

In addition, the known method for monitoring defective areas in lumber, which includes comparison with a reference color and performing binary conversion, does not include accurate defect monitoring using color distribution.

The aim of the present invention is to solve these problems of the prior art and to provide the ability to accurately detect defective areas formed due to changes in the surface color of lumber, which affect the quality of the lumber, using the color distribution from the captured image of the lumber, for example veneer captured by the imaging tool, even In the event that there are changes in color, brightness, etc. abnormal area.

Problem Solving Tools

In FIG. 1 shows an exemplary layout of a veneer sorter. In FIG. 1, reference numeral 1 denotes an image processor (image processing means), 2 denotes a sorter controller, 3 denotes an operation panel, 4 denotes a conveyor belt, 5 denotes an irradiation for transmitted light, 6 denotes an irradiation for reflected light, 7 denotes a sorting device, 8 denotes a camera with a single-line sensor (imaging tool), and 9 denotes veneer (lumber).

To solve the above problems, the present invention is configured as follows.

(1) The image forming means 8 captures the color image of the lumber 9. The image processing means 1 obtains the color distribution of the color image captured by the image forming means 8, compares the obtained color distribution with a predetermined color distribution for normal lumber, evaluates the obtained color distribution as abnormal in the case its deviations from the color distribution of normal lumber by a predetermined value or more and determine it possible in the timber as a defect area on the surface of the timber, the captured image forming means 8, where the abnormal color distribution exceeds a predetermined value. Thus, using the color distribution, it is possible to accurately detect defective areas formed due to color changes in the surface of the lumber, which affect the quality of the lumber.

(2) In the lumber inspection method described above or the device according to (1) when lumber is installed as an inspection object in which the area of the defective zone is small relative to the area of the inspection object, the image distributions obtained for each object are used interchangeably as predetermined color distributions of normal lumber. Therefore, it is possible to easily obtain predetermined color distributions of normal lumber.

(3) In the above lumber control method or device according to (1) or (2), an abnormal brightness portion is detected by obtaining a brightness histogram for the color distribution of the captured color image. In this way, an abnormally bright area, such as a burn, can be easily detected.

(4) In the above lumber control method or device according to (3), it is assumed that the full brightness histogram of a predetermined color distribution for normal lumber corresponds to the normal distribution, and the total normal distribution is estimated based on the cumulative frequency of the partial region. Consequently, the color distribution for normal lumber can be estimated based on the controlled lumber to be controlled, without the need for a preliminary determination of the color distribution of normal lumber.

Advantages of the Invention

The present invention has the following beneficial effects.

(1) The image processing means obtains the color distribution of the color image captured by the image forming means, compares the obtained color distribution with a predetermined color distribution for normal lumber, evaluates the obtained color distribution as an abnormal distribution when it deviates from the color distribution of normal lumber by a predetermined amount, or more and determines as a defect of lumber the zone on the surface of the lumber, s hvachennuyu imaging means, wherein the abnormal color distribution exceeds a predetermined value. Thus, using the color distribution, it is possible to accurately detect defective areas formed due to color changes in the surface of the lumber, which affect the quality of the lumber.

(2) When installing sawn timber as an inspection object, in which the defective area is small in relation to the area of the inspection object, the image distributions obtained for each inspection object are used interchangeably as predetermined color distributions of normal lumber. Therefore, it is possible to easily obtain predetermined color distributions of normal lumber.

(3) Since an abnormal brightness portion is detected by obtaining a brightness histogram for the color distribution of the captured color image, an abnormal brightness portion such as a burn can be easily detected.

(4) It is assumed that the full brightness histogram of the predetermined color distribution for normal lumber corresponds to the normal distribution, and the total normal distribution is estimated based on the cumulative frequency of the partial region. Therefore, the color distribution for normal lumber can be estimated based on the lumber to be controlled, without first having to determine the color distribution of normal lumber.

Brief Description of the Drawings

FIG. 1 is an exemplary diagram of a veneer sorter according to the present invention;

FIG. 2 is an exemplary diagram of an image processor according to the present invention;

FIG. 3 is an exemplary diagram showing color conversion of corresponding image points on an HS plane according to the present invention;

FIG. 4 is an exemplary block diagram of a relaxation according to the present invention;

FIG. 5 is an exemplary color distribution diagram in an HSC color space according to the present invention;

FIG. 6 is an exemplary pixel distribution diagram on a particular brightness plane v according to the present invention;

FIG. 7 is an exemplary diagram of the cumulative frequency F (x) of the normal distribution according to the present invention;

FIG. 8 is an exemplary pixel distribution diagram (normal distribution profile) in the direction of the central axis according to the present invention;

FIG. 9 is an exemplary diagram for estimating an average Vm according to the present invention;

FIG. 10 is an exemplary diagram for a case where an abnormal brightness portion occupies a large area according to the present invention.

Character Description

1 denotes an image processor (image processing means);

2 denotes a sorter controller;

3 indicates an operation panel;

4 denotes a conveyor belt;

5 denotes radiation for transmitted light (irradiation means);

6 denotes radiation for reflected light (irradiation means);

7 denotes a sorting device;

8 denotes a camera with a single-line sensor (image forming means);

9 denotes veneer (lumber).

Best Mode for Carrying Out the Invention

(1) Veneer Sorter Description

In FIG. 1 shows an exemplary layout of a veneer sorter. This figure shows the complete configuration of a veneer sorter. The veneer sorter includes an image processor 1, a sorter controller 2, an operation panel 3, a conveyor belt 4, an irradiation 5 for transmitted light, an irradiation 6 for reflected light, a sorting device 7, and a camera 8 with a single-line sensor.

The image processor 1 processes the image data obtained from the camera 8 with a single-line sensor and outputs the processing results, for example, veneer quality class, to the sorter controller 2. The sorter controller 2 provides control signals for the sorter conveyor, for example, to drive and stop the conveyor, as well as control signals for the sorting device 7 in response to the output signal of the image processor 1. The operation panel 3 is designed to perform operations such as changing the settings of the image processor 1 and controlling the controller 2 of the sorter. The conveyor belt 4 carries out the transportation of veneer 9. The irradiation 5 for transmitted light is an irradiating means (light source), for example a light emitting diode (LED), for detecting holes, cracks, and the like. in veneer 9, moreover, its irradiation differs in color (for example, green radiation) from irradiation 6 for reflected light. This is done in order to distinguish (by color and intensity) the transmitted light from the reflected light coming from the irradiating means 6 for reflected light, for detecting holes (from knots), cracks, etc. in the veneer. Irradiation 6 for reflected light is an irradiating means (light source), such as LED, for detecting light reflected from veneer 9, and white radiation is usually used. A camera 8 with a single-line sensor captures line images of veneer 9.

The operation of the veneer sorter includes capturing the image of the veneer 9 transported by the conveyor belt 4 using the camera 8 with a single line sensor, and outputting the image data to the image processor 1. The image processor 1 processes the image data and outputs to the sorter controller 2 processing results, for example, a veneer quality class. The sorter controller 2 outputs a control signal to the sorting device 7 and sorts the plywood 9 according to the quality classification. Sorting is performed in accordance with the number of wormholes, holes or knots that have fallen, fresh knots, dried knots, pickles, cracks, pitches and pockets with bark, blue rot, deformations (warping), etc., and also in accordance with their sizes.

(2) Description of the image process

In FIG. 2 is an exemplary diagram of an image processor 1. In FIG. 2, the image processor 1 includes three cameras 8a, 8b, and 8c with a single line sensor; camera image panels 11a, 11b and 11c; laser markers 12a and 12b; laser exciters 13a and 13b and the host computer 14.

Cameras 8a, 8b and 8c with a single-line sensor share the veneer 9 in the direction perpendicular to the direction of transportation between the three cameras and capture color images. The camera image panels 11a, 11b, and 11c each time digitize a single-line image loaded from the respective cameras with a single-line sensor and transmit the image data to the main computer 14. Laser markers 12a and 12b emit thin light beams in the direction of veneer transport, which are intended for use as marks in the synthesis (connection) of the corresponding images from cameras 8a, 8b and 8c with a single-line sensor. The emitted light beam may be a thin light beam of a different color (for example, a monochromatic light beam of a red laser) depending on the color of the veneer (lumber) so that it can be easily removed in subsequent processes. Laser exciters 13a and 13b are connected to an AC power source and excite laser markers 12a and 12b. The main computer 14 includes processing means, memory means, output means, and the like. and performs image processing (image synthesis, knot search, defect search, etc.) of veneer 9. In this case, cameras 11a, 11b and 11c of receiving images from cameras and main computer 14 form image processing means.

The operation of the image processor includes irradiating the conveyed veneer 9 with light from the irradiation means 5 for transmitted light and the irradiation means 6 for reflected light. Each time a single-line image is loaded from cameras 8a, 8b and 8c with a single-line sensor, the image acquisition panels 11a, 11b and 11c transmit their data to the main computer 14. The main computer 14 compensates the received images and determines the amount of warpage, and then sequentially connects images. As a result, when the loading of images on the image acquisition panels 11a, 11b and 11c ends, the main computer 14 almost completes the synthesis of the color image and the conversion of the black and white grayscale image. Then, the main computer 14 connects the images from the cameras 11a, 11b and 11c of the image acquisition from the cameras, where the veneer image is divided into three parts.

At this point, the laser markers 12a and 12b emit laser marks on the veneer 9 to divide the veneer 9 into three parts. In addition, by aligning the line images with the corresponding laser marks on the veneer 9 provided by the cameras 8a, 8b and 8c with a single line sensor, the image can be easily connected. In addition, to increase the speed of image processing, the search for knots can be performed using black and white halftone images that have a large number of pixels, while searching for dried knots, etc. can be done using color images at a reduced scale (with a reduced number of pixels).

Next, in two stages, the operation of the image processor will be described, namely during image capture and subsequent processing.

<Description of processing during image capture>

Image data captured by cameras 8a, 8b, and 8c is delivered line by line to main computer 14 and synthesized as one common image.

The image acquisition panels 11a, 11b and 11c of the cameras download a single-line color image from cameras 8a, 8b and 8c with a single-line sensor, detect the position of the laser mark (connection position) and send the single-line color image along with information about the position of the laser mark to the main computer 14.

The main computer 14 compensates for the incoming single-line color image, simultaneously determines the “warping amount” and performs synthesis based on the above position information (laser mark). At this moment, if there is warpage in the controlled object, then a non-linear distortion of the trajectory appears at the place of the laser mark, as a result of which, having estimated the distortion value, it is possible to determine the distortion value.

At the stage when the image capturing is completed by the camera image panels 11a, 11b and 11c and the last single-line color image is received, the main computer 14 completes the synthesis of the entire color image. In addition, for efficient use of time during image capture, processing performed for each individual line, such as black-and-white conversion and downsizing, can be performed simultaneously.

<Description of processing during image analysis after capture>

Panels 11a, 11b and 11c for receiving images from cameras are in a waiting state until the arrival of the next board (lumber) is detected.

The host computer 14 based on predefined information such as size, type or the like. the board being the object of control performs a search for knots, detects defects using transmitted light, etc. in accordance with the evaluated zone and the settings, and, in the end, performs sorting, including the determination of the warpage value. Sorting results are displayed (not shown), and also displayed on the controller 2 of the sorter.

Although the above description relates to the case when computers (personal computers), a main computer 14, and the like are used as panels 11a, 11b, and 11c for receiving images from cameras, in the image processor, the number of computers used may vary depending on the amount of image data, computer performance, or the like. Alternatively, processing may be performed by one computer.

In addition, although the case of using three cameras 8a, 8b and 8c with a single-line sensor is described here, in the alternative, one, two or four or more cameras with a single-line sensor can be used depending on the size and type of board being monitored, computer performance, or t .P.

(3) Description of the detection of a defective area resulting from a change in the color of the surface of the veneer

Defective areas formed due to color changes in the surface of the lumber, which affect the quality of the lumber, can be detected using the tools described below according to the method described below.

A defective area formed as a result of a color change refers to: a site whose color has changed under the influence of a fungus (or the like) that has fallen on the lumber from the outside; or in the case of veneers used in glued plywood or the like, to a burn formed in the dryer; to the track on the bark of lumber; to the resin inside the lumber and leaked to its surface, etc.

Now such areas can be diagnosed as defective areas.

(Description of defective area detection means)

1) Cameras 8a, 8b and 8c with a single-line sensor capture an image of the surface of the lumber, which is downloaded to the computer using an input device (input means).

2) The captured image is saved as a color image in which each pixel is composed of a combination of RGB (red, green and blue) in the computer's memory (storage medium).

3) A computer image processing program (image processor) converts an RGB image into an HSV image (hue, saturation, and brightness).

4) A defective portion is detected based on the HSV image as described below.

(Description of the method for detecting a defective area)

The surface color of the normal (“healthy”) lumber in the same samples is distributed in a specific area of saturation and tone, regardless of whether it is dark or light. However, a defective area, for example, affected by a fungus, has a deviation from the color distribution in a healthy area both in saturation and in tone due to the difference in their nature. In addition, the color distribution of the defective area, for example, with a burn, is located in a much darker area (with lower brightness) compared to the color distribution of the healthy area.

Thus, this method is characterized in that it reveals deviations of saturation and tone, as well as a deviation of the brightness of the color distribution of the surface of the controlled lumber compared to the color distribution of the surface of healthy lumber and identifies a section with a significant deviation as a defective section.

(4) Description of a method for obtaining a color distribution of the surface of healthy lumber to be used as a reference

1) A color camera with a single-line sensor captures the surface image of healthy lumber for control specimens.

2) In step 1), to ensure sufficient statistical accuracy for the same specimens, a set (of the order of 20 or more) surface images is taken under various conditions.

3) Each pixel color in all of the above images is arranged in a three-dimensional color position in computer memory to create a three-dimensional color distribution.

As for the three-dimensional color space, you can use the RGB color position (red, green, and blue), the HSV color space (hue, saturation, and brightness), the Lab space (where “L” stands for brightness, “a” stands for tone element in the range from green to red, and “b” refers to a tone element in the range from blue to yellow) or the like.

4) For each plane with the same brightness in a three-dimensional color distribution, a two-dimensional distribution is obtained to determine a point indicating the maximum frequency.

5) By gradually changing the brightness, you can get a curve that approximately connects the maximum frequency points in 4). This curve is called the reference central axis of the three-dimensional color distribution.

For example, when the brightness value is in the range from 0.0 to 1.0 in the HSV color distribution, for each 0.02 brightness segment, a two-dimensional color distribution of the hue and saturation of pixels having the same brightness value is obtained, and a curve connecting the maximum frequency, which is set as the reference Central axis of the three-dimensional color distribution. At the same time, the standard deviation σc (v) of the two-dimensional distribution of tone and saturation is obtained.

6) When it is known in advance that the area of the defective area is small with respect to the area of the control object, the reference distribution can be replaced by the image distribution obtained for each control object. This is possible because it is necessary to obtain only the average distribution value for a healthy site and the standard deviation value.

(5) Description of defect inspection method

1) Using a color camera with a single-line sensor, they capture the image of the surface of the controlled lumber.

2) Each image pixel is positioned in a three-dimensional color space to create a three-dimensional color distribution.

3) Get a chromatic deviation from the reference Central axis of the three-dimensional color distribution as follows.

For example, if the pixel at position x, y in the image of the object is designated as d [x, y], then the color in the HSV color space is expressed as

tone value: h (g [x, y]),

saturation value: s (g [x, y]),

brightness value: v (g [x, y]),

the coordinates of the reference central axis at a certain brightness v on the reference central axis of the three-dimensional color distribution obtained earlier are expressed as

tone value: H (v), saturation value: S (v), and the abscissa and ordinate of a plane with the same brightness are denoted as X and Y, respectively; after which we obtain a diagram, FIG. 3. In FIG. 3 is an exemplary diagram showing the color conversion of the corresponding pixels of the image g onto the HS plane. In FIG. 3, the pixel g [x, y] of the image of the object is converted to the Cartesian coordinates X2, Y2 on the HS plane. Although the color distribution (see the shaded area) may take different forms from the circle, here we get an approximate circle with a standard deviation.

Cartesian coordinates X1, Y1 of the coordinates H (v), S (v) of the reference central axis can be represented here as

X1 = S (v) cos (2π H (v) / 360),

Y1 = S (v) sin (2πH (v) / 360).

Cartesian coordinates X2, Y2 of the pixel g [x, y] can be represented as

X2 = S (v) cos (2π h (v) / 360),

Y2 = S (v) sin (2πh (v) / 360).

The square of the spatial distance from the reference central axis can be obtained in the form

d2 = (X1 · X2) ² + (Y1 · Y2) ² .

Thus, the chromatic deviation Zc [x, y] can be expressed as

Zc [x, y] = (√d) / (σc (v) × βc),

where σc (v) denotes the standard deviation of the two-dimensional color distribution of hue and saturation at brightness v along the reference central axis, and βc denotes a coefficient for determining how much the color should deviate from the reference central axis in order to consider it anomalous, which is estimated by the σc ( v) which, for example, take a value from about 1.0 to 2.0.

In addition, the spatial distance deviation can be similarly obtained using other color distributions, for example, Lab color space.

4) Further, in order to calculate the actual area of the defective area, it is necessary to select only those color pixels that deviate from the standard in the color distribution space, where the original pixel g [x, y] of the lumber locally forms a cluster. In this case, you can use a method called relaxation, which is a common image processing technology, to remove individual isolated points and highlight areas with a significant deviation taking into account the colors of the surrounding pixels.

As an example, a defective zone can be determined due to an anomalous color by relaxation, in which the value of the chromatic deviation Zc [x, y] from the reference central axis is used as the initial mark (see the description of Fig. 4).

(Description of detection of a defective area by anomalous brightness)

(5) We obtain a histogram in the direction of the reference central axis (brightness axis) for a three-dimensional color distribution.

(6) If the brightness histogram of the healthy section described above has a normal distribution (Gaussian distribution) with an average value of Vm and a standard deviation σv, then the value of the brightness deviation Zv [x, y] can be obtained as

Zv [x, y] = | Vm-g [x, y] .v | / (σv × βv),

where βv denotes a coefficient for determining how much color should deviate from the average brightness value Vm so that it can be considered anomalous, which is estimated by the indicator σv, which takes a value from about 1.0 to 4.0.

The total value Zt [x, y] of the deviation of color and brightness can be expressed as

Zt [x, y] = Zc [x, y] + Zv [x, y].

If the brightness histogram does not form a normal distribution due to the characteristics of the camera or the like. (for example, when saturation appears in a bright region close to luminance 1.0), an accurate calculation of the average value of Vm and the standard deviation σv may not be necessary. In this case, if we assume that the histogram obeys the standard normal distribution, the normalized function F (x) of the cumulative probability distribution can be expressed as

[Expression 1]

where x is luminance, µ is the average luminance, and σ is the standard deviation.

The luminance histograms are integrated, starting from the minimum luminance, and luminances are obtained that are designated as V1, V2, V3 and Vm, the value of which (cumulative frequency) obtained by dividing the integrated value by the total number of pixels corresponds to p1, p2, p3 and p4 shown below.

p1 = F (µ-2.0σ) = 0.0228

p2 = F (µ-1,0σ) = 0.1587

p3 = F (µ-0.5σ) = 0.3085

p4 = F (µ) = 0.5

In addition, the allowable ranges Vmin and Vmax achievable with V1, V2, V3 and Vm can be obtained empirically from lumber or the like used as a reference. For example, a possible setting may be Vmin = 0.25; Vmax = 0.9.

a) If the brightness position corresponding to V2: F (Vm-σ) = 0.1587 and Vm: F (Vm) = 0.5, which is in the acceptable range, can be found among V1, V2, V3 and Vm, then the estimated the mean value x = Vm, and the standard deviation σv can be obtained based on FIG. 8.

b) However, in the case where the distribution profile is violated in the area where the brightness distribution is less than x = µ, then Vm falls outside the allowable range of the region. In this case, by calculating V1: F (Vm-2.0σv) and V2: F (Vm-1.0σv), which are within the acceptable range, estimates of σv and Vm can be obtained as

σv = V2-V1

Vm = V2 + σv

(see Fig. 9).

c) Alternatively, when the area with anomalous brightness becomes relatively large, V1: F (Vm-2.0σv) and V2: F (Vm-1.0σv) fall out of the acceptable range. In this case, based on the brightness values V3: F (Vm-0.5σ) = 0.3085 and Vm: F (Vm) -0.5, which are in the acceptable range, σv can be expressed as σv = (Vm-V3) × 2, which allows us to estimate the values of σv and Vm (see Fig. 10).

Using this method, you can get the average value and standard deviation for a healthy site without the need for lumber, which serves as a reference, and regardless of the distribution profile. When using lumber serving as a reference, the average value and standard deviation for it should be used without performing calculations.

7) Next, to calculate the actual area of the defective area, it is necessary to select only those color pixels that deviate from the standard in the color distribution space, where the original pixel g [x, y] of the lumber locally forms a cluster. Moreover, to remove individual isolated points and highlight areas with a significant deviation, you can use a method called relaxation, which refers to the general image processing technology that takes into account the colors of the surrounding pixels.

For example, a defective area due to abnormal color and luminance can be determined using relaxation, which uses the total value Zt [x, y] of the deviation of color and luminance as an initial mark (see description of FIG. 4).

Previously, in automatic quality control of lumber, defective areas were detected only by how light or dark the color of the surface, or by determining a specific color. However, these methods are unable to perceive changes in the brightness and color of the plots, which leads to the assessment of all controlled objects as defect-free.

Areas that look unpleasant to the human eye and that affect the quality of the material are likely to have colors that are unnatural for lumber. These colors appear as deviations in the color distribution in the three-dimensional color space. Defective areas can now be detected with high accuracy by using a unified method in which these differences are distinguished and diagnosed.

In addition, since a fungus and the like, affecting the quality of lumber, has different colors depending on the type or area of production of lumber, it is difficult, using one method, to find all of these options for fungi with high accuracy. However, according to the present invention, even if the types of lumber as objects of control are different, a defect can be detected without a strong change in the reference values of the coordinates of the central axis. Even if the accuracy of detection deteriorates depending on a particular instance, the accuracy of detection can be restored simply by changing the coordinates of the reference central axis of the original three-dimensional color distribution.

In addition, although resin, bark, etc. it is difficult to detect by visual inspection using image processing, now, having the above method, such defects can be detected.

Because the darkened areas of the cortex, etc. can now be detected with high accuracy, it is easy to assess whether this knot is alive or dry, by assessing whether the darkened bark remains or not.

(6) Description of relaxation

In FIG. 4 is an exemplary relaxation diagram. The following describes the processes S1 to S3 shown in FIG. four.

In these processes, for each pixel g (x, y) in the image that becomes the object of control, the probability Pi (x, y) is set, where Pi (x, y) is the probability of a pixel defect g (x, y) after the ith iterations.

S1: The image processor assigns to each pixel g (x, y) in the image that becomes the object of control, the initial probability P0 (x, y) (from 0 to 1.0) and proceeds to process S2. In process S1, the probability P0 (x, y) is determined as follows using the color deviation Z (x, y).

P0 (x, y) = Z (x, y): 0 <Z <1,0

1.0: Z≥1.0

S2: For all pixels,

if (0,0 <Pi (x, y) <1,0),

then the image processor calculates the average probability value <P> of neighboring pixels Pi (x, y), updates this probability as Pi + 1 = Pi + α (<P> -Pi) (where α is the influence coefficient of neighboring pixels that are in range from about 1 to 4) and goes to process S3.

If (Pi (x, y) ≤0.0 or Pi (x, y) ≥1.0), then the image processor assumes that

Pi + 1 = Pi

and proceeds to process S3 without updating the probability.

S3: The image processor checks the convergence condition.

If for Pi (x, y) the number of iterations I is more than a given number (I> a given number) and the percentage of pixels with Pi = 0,0 and Pi = 1,0 among all pixels is more than a given percentage (> a given percentage), then this the process ends.

If not, then process S2 is repeated.

In this case, the set number of iterations is set to about 10, and the set percentage of pixels with Pi = 0.0 and Pi = 1.0 among all pixels is set to about 99%.

(7) Description of the color distribution in the HSV color space

In FIG. 5 shows an exemplary color distribution diagram in the HSV color space. In FIG. 5 in the vertical direction represents the brightness (V: in this case, V = from 0.0 to 1.0), in the diameter direction on the same brightness plane, the saturation is presented (S: in this case, S = from 0.0 to 1, 0), and in the circumferential direction on the same brightness plane, the tone is represented (H: in this case, H = from 0º to 360º). The color distribution for healthy lumber includes significant upper and lower areas of color distribution. The central axis (reference central axis) is indicated by an arrow pointing up.

In addition, the color distribution in the area with anomalous color, which is a discolored portion, such as a fungus, is shown on the right side as an area of anomalous color. In addition, in the lower part, in the form of a small region of color distribution, an abnormal brightness portion is shown, for example, a burn caused by treatment in a dryer or the like.

(8) Description of the pixel distribution in a specific brightness plane

In FIG. 6 is an explanatory diagram of a pixel distribution in a specific brightness plane v. In FIG. 6 shows the pixel distribution for a region with anomalous color in a particular brightness plane v. In this case, the pixel distribution for the area with anomalous color occupies a specific area (see the shaded area) with a standard deviation σс (v).

(9) Description of the pixel distribution in the direction of the central axis

In FIG. 7 is an exemplary cumulative frequency diagram F (x) of a normal distribution. In FIG. 7, the total normal distribution (probability of distribution) is shown by a dashed line, and the cumulative frequency (cumulative probability distribution function) F (x) is normalized by dividing the integral value for a healthy section of lumber by the total number of pixels N. In this case, p1 = 0.0228 (µ -2σ), p2 = 0.1587 (μ-1.0σ), p3 = 0.3085 (μ-0.5σ) and p4 = 0.5 (average value = μ).

In FIG. 8 is an explanatory diagram of a pixel distribution (normal distribution profile) in the direction of the central axis. In FIG. 8 shows a histogram in the direction of the reference central axis (luminance axis) of the three-dimensional color distribution. On the luminance histogram, the luminance values corresponding to p1, p2, p3 and p4 described above are V1, V2, V3 and Vm, respectively, and the acceptable limits of the range attainable on a healthy section of the sawn timber surface are Vmin and Vmax. In this case, if the brightness position corresponding to V2: F (Vm-σ) = 0.1587 and Vm: F (Vm) = 0.5, which are in the acceptable range, can be found, then the average value will be x = Vm and one can get the standard deviation σv.

In FIG. 9 is an explanatory diagram for estimating the average value of Vm in the case of an abnormal distribution profile. In FIG. 9 shows a histogram in the direction of the reference central axis of the three-dimensional color distribution. With such a histogram of brightness, in the case where the distribution profile is violated in the area where the brightness distribution of the healthy section of the sawn timber surface is less than x = µ, then Vm falls outside the allowable range. In this case, σv and Vm can be estimated by calculating V1: F (Vm-2,0σv) and V2: F (Vm-1,0σv), with σv = V2-V1, and Vm = V2 + σv.

In FIG. 10 is an exemplary diagram for the case where the area with anomalous brightness occupies a large area. In FIG. 10 shows a histogram in the direction of the reference central axis (luminance axis) of the three-dimensional color distribution. With such a histogram of brightness, in the case when the area of the section with anomalous brightness on the surface of the lumber is relatively large, then V1: F (Vm-2.0σv) and V2: F (Vm-1.0σv) fall out of the allowable range. In this case, based on the brightness values V3: F (Vm-0.5σ) = 0.3085 and Vm: F (Vm) = 0.5, which are in the acceptable range, the value of σv can be expressed as

σv = (Vm-V3) × 2, which allows us to estimate σv and Vm.

As shown here, when the brightness histogram does not have a normal distribution shape due to the characteristics of the camera or the like. (for example, when the sensitivity characteristics of the sensors are nonlinear in a region close to a brightness of 1.0), it is not necessary to obtain the exact average value of Vm and the standard deviation σv. However, even in this case, if we assume that the histogram corresponds to the standard normal distribution, the total distribution can be estimated from the histogram on the decay of the distribution (two points), we can obtain the average value of Vm and the standard deviation σv.

Accordingly, the average value of Vm and the standard deviation σv for a healthy site can be obtained (estimated) using an image processor using two points from the number of points from V1 to Vm that are in the acceptable range (from Vmin to Vmax). In order of priority of the indicated two used points, the average value Vm and the other point must first be used (one point from the number of points V1 to V3), and two points from the points V1 to V3 should be used if the average value of Vm is not in the allowable range .

(10) Description of program installation

Image processor 1 (image processing means), sorter controller 2 (sorter control means), camera image receiving panels 11a, 11b and 11c, main computer 14, and the like. can be configured as a program executed by the central processing unit (CPU) and stored in the main storage. This program must be processed by a computer. The computer is made of hardware, such as a main control unit, main memory, a file device, an output device, such as a display, and an input device.

The program according to the present invention is installed on said computer. Installation is performed by first storing the program on a portable recording medium such as a floppy disk, magneto-optical disk, etc., and installing the recording medium on a file device included in the computer, or through the drive provided on the computer to provide access to the recording medium , or through a network, such as a local area network (LAN). Accordingly, it is possible to easily provide a lumber control device capable of accurately detecting defective areas using color distribution due to a change in the surface color of the lumber that affects the quality of the lumber.

Claims (7)

1. The method of control of lumber, characterized in that the method comprises the steps in which:
using image forming means to capture a color image of the lumber; and
use the image processing means to obtain the color distribution of the color image captured by the image forming means, compare the brightness histogram of the obtained color distribution with a predetermined histogram of the brightness of the color distribution of normal lumber, evaluate the resulting color distribution as abnormal when the brightness histogram of the obtained color distribution deviates from the predetermined histogram by a predetermined amount or more, and about detecting a zone on the surface of the lumber captured by the image forming means whose abnormal color distribution has a value exceeding a predetermined value as a lumber defect. 2. The method of controlling lumber according to claim 1, characterized in that the color distributions obtained from the color image of normal lumber are used from time to time as predetermined color distributions of normal lumber. 3. The method of controlling lumber according to claim 1 or 2, characterized in that for the histogram of the brightness of the color distribution of the normal lumber determined in advance, the total normal distribution is estimated based on the cumulative frequency of the partial region under the assumption that the total distribution follows the normal distribution. 4. A device for controlling lumber, characterized in that the device contains:
image forming means for capturing a color image of lumber; and
image processing means for obtaining a color distribution of a color image captured by the image forming means, comparing a luminance histogram of a obtained color distribution with a predetermined luminance histogram of a normal lumber color distribution, evaluating the obtained color distribution as abnormal when the luminance histogram of the obtained color distribution deviates from a predetermined luminance histogram by a predetermined amount or more, and the zone on the surface of the lumber captured by the image forming means, whose anomalous color distribution has a value that exceeds a predetermined value, as a lumber defect. 5. The lumber control device according to claim 4, characterized in that when installing the lumber, whose defective area is small in relation to the area of the control object, the image processing means uses color distributions obtained from the color image of the normal part of the lumber as predefined color distributions of normal lumber. 6. The lumber control device according to claim 4, characterized in that for the brightness histogram, the image processing means estimates the total normal distribution based on the cumulative frequency of the partial region under the assumption that the full distribution corresponds to the normal distribution. 7. A computer-readable storage medium having a program stored on it which, when executed on a computer, instructs it to implement a method comprising the steps of:
using image forming means to capture a color image of the lumber; and
use the image processing means to obtain the color distribution of the color image captured by the image forming means, compare the brightness histogram of the obtained color distribution with a predetermined histogram of the brightness of the color distribution of normal lumber, evaluate the resulting color distribution as abnormal when the brightness histogram of the obtained color distribution deviates from the predetermined histogram brightness by a predetermined value or more and detecting a zone on the surface of the lumber captured by the image forming means whose abnormal color distribution has a value that exceeds a predetermined value as a lumber defect.

Images