JPH0735532A - Method of measuring length of crack on road surface - Google Patents

Abstract

PURPOSE:To measure a length of a crack in a unit area by setting a unit area in a binary image indicating a crack, and extracting NXN pixels about the pixel thereof, and by comparing the pixel array pattern with a preobtained pixel array pattern so as to obtain a crack length factor. CONSTITUTION:A pixel P1 indicates a part which cracks, and an array patter of 3X3 pixels about the pixel P1 is compared with a plurality of pixel array patterns. A crack length factor with which the patterns are coincident with each other is stored in a suitable memory position. A pixel P1(1) indicates a part which cracks, and accordingly, an array pattern of pixel surrounding the pixel is compared with a plurality of patterns. A crack length pattern allocated to the coincident patter is obtained, and is added to the previously obtained factor. A pixel P1(2) indicates a part which does not crack, and accordingly, is neglected. Similarly, pixels P2(1), P2(2), P3(2) are compared with the array patterns, and thus obtained factors are added. From thus obtained added factor, the length of a crack can be obtained.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for measuring the length of cracks on a road surface, and more particularly, to a binary image representing the presence or absence of cracks by performing image processing on an analog image of the road surface. To a measurement method for obtaining the crack length of the road surface per unit area from.

[0002]

2. Description of the Related Art "Cracks" on a road surface in order to evaluate the condition of the paved road surface or to specify the repair time of the paved road surface.
There is a method to measure the length of. Conventionally, a road surface is photographed to obtain an analog image, the analog image is subjected to image processing to obtain a binary image representing a crack on the road surface, and the crack length of the road surface is obtained from the binary image. ing.

The road surface on which the length of the crack is to be measured is usually photographed using a long film, but it may be photographed using a video camera. When shooting a road surface using a video camera, a method is known in which an analog image is digitized by a video digitizer after shooting or the output of the video camera is directly digitized. The present invention is not limited to the road surface photographing method, and can be applied to a binary image obtained using any of the above methods.

Prior to the description of the embodiments of the present invention, a conventional image processing method for obtaining a binary image by subjecting an analog image of a road surface to image processing to obtain a binary image and obtaining the crack length of the road surface from the binary image Will be briefly described.

First, the road surface is photographed so as to include the center line, and an analog image (not shown) is created.

(A) From this analog image, a digital image (image consisting of multi-valued pixels) shown in FIG. 13 is obtained using a known method.

(B) The noise caused by the minute unevenness of the road surface included in the multi-valued pixel digital image of FIG. 13 is removed by a known method (method using a low-pass filter), and the noise shown in FIG.
The digital image shown in 4 is obtained.

(C) A known Sobel operator is used to detect edges from the digital image shown in FIG. 14 to obtain the digital image shown in FIG.

(D) A binary image is obtained from the digital image shown in FIG. For example, when each pixel in FIG. 15 is an 8-bit multi-valued pixel, the binary pixel image shown in FIG. 16 is obtained by setting the threshold value for binarization to 16.

(E) In order to reduce the data amount of the binary image of FIG. 16, 5 × 5 pixels (25 pixels in total) of the binary image of FIG. 16 are set as a new pixel. For example, if the number of pixels having a value of 1 (representing a cracked portion) among 25 pixels is 12 or more, the new pixel is set to 1 (represents a cracked portion), and if the number of pixels is less than 12, If there is, the new pixel is set to 0 (indicating that it is not a cracked portion). The obtained binary image is shown in FIG.

(F) Further, the missing points missing during the data amount reduction processing in step (e) are interpolated. That is, 3 × 3 pixels (9 pixels in total) are sequentially extracted from the pixel image of FIG. 17 while shifting the center pixel by one pixel.
When the value of the center pixel of the extracted 3 × 3 pixels is 0, and if, for example, 4 or more pixels are 1 in the 8 surrounding pixels, the value of the center pixel is replaced with 1, and the pixel in the 8 surrounding pixels Number of 1 is 4
If it is less than the number, the value of the central pixel is left at 0. FIG. 18 shows a binary image in which missing points are interpolated.

(G) Finally, isolated point noise is removed from the image of FIG. That is, in the area of 5 × 5 pixels (total of 25 pixels), when the values of all 16 pixels of the outermost circumference are 0,
The values of the inner 9 pixels are all set to 0. Next, in the area of 7 × 7 pixels (total of 49 pixels), when the values of the outermost 24 pixels are all 0, the values of the inner 25 pixels are all 0. Such image processing is performed for each pixel, and
The binary image shown in is obtained.

The crack length of the road surface is obtained from the binary image (FIG. 19) obtained by the image processing of the above (a) to (g).

With the center line in the binary image of FIG. 19 as a reference, the binary image of FIG. 19 is divided into a plurality of square unit pixel areas (for example, a square pixel block of which one side is 50 cm). The crack length in each pixel block of a plurality of unit areas is measured, and the total crack length of each block is output as a measurement result. Since the binary image of FIG. 19 is divided into a plurality of pixel blocks with the center line of the road surface as a reference, a pixel block that is less than the unit area occurs in a portion corresponding to the end of the road.
In this case, after measuring the crack length within the pixel block which does not reach the unit area, this measured value may be converted into the crack length of the unit area. For example, if the area of the pixel block at the end of the road is 1/2 of the pixel block of the unit area, the crack length of the pixel block at the end of the road is doubled.

By the way, the conventional method for obtaining the crack length in each pixel block is to trace the cracked portion in the pixel block as a line.

However, the conventional method of tracing the cracked portion as a line has the following problems. That is, (a) it takes a considerable amount of time for the thinning process to make it easier to trace cracks as lines. (b) If the cracks are reticulated, a very long processing time is required. (c) Many branches (judgments) to trace cracks as lines
Due to the existence of conditions, it takes a huge amount of processing time to measure the crack length.

For this reason, there has been a demand for a measuring method which significantly reduces the processing time without using the conventional method of tracing the crack length as a line.

[0018]

SUMMARY OF THE INVENTION An object of the present invention is to determine the crack length of a road surface in a very short processing time without using the conventional method of tracing a crack as a line.

[0019]

According to the present invention, a road surface is photographed to obtain an analog image of the road surface, and the analog image is subjected to image processing to obtain a binary image representing the presence or absence of cracks.
Regarding a method for obtaining a crack length per unit area of a road surface from a value image, (a) a plurality of pixel array patterns which can be binary N × N pixels (N is an odd number of 3 or more) of which the central pixel indicates a crack Is calculated in advance, and a coefficient of the length of the crack contributed by the central pixel representing the crack is calculated in advance for each of the plurality of pixel array patterns, and (b) a unit area is set in the binary image representing the crack. (C) N out of the pixels in the unit area, with the pixel representing the crack being the center
XN pixels are extracted, and (d) the extracted pixel array pattern of N × N pixels is compared with a plurality of pixel array patterns obtained in advance to obtain the same pixel array pattern, which is given to the pixel array pattern. Calculate the coefficient of crack length
(E) The steps (c) to (d) are repeated for each pixel in the unit area, and the coefficients obtained respectively are added,
A method for measuring a crack length of a road surface, characterized in that the crack length in the unit area is measured based on the sum of the obtained coefficients.

[0020]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to FIGS. 1 to 12 and 19. The image processing until obtaining the binary image shown in FIG. 19 is the same as the conventional one.

It can usually be assumed that the width of the crack is sufficiently narrow compared to its length. Therefore, FIG.
It is possible to approximate 1/2 of the length of the contour line of the cracked portion (shown in black in the drawing) as the cracked length. That is, the total of 1/2 of the length of the contour line of the cracked portion can be obtained without following the crack.

In the present embodiment, the binary image of FIG. 19 is divided into unit areas of a square having a side of 50 cm with the center line as a reference, and each pixel in each unit area is sequentially arranged, as in the conventional case. 0 (not belonging to the cracked portion (shown in white)) or 1 (belonging to the cracked portion (shown in black))
To judge. If the pixel is 0, ignore that pixel,
When the number of pixels is 1, the pixel array pattern consisting of the surrounding 8 pixels is compared with a plurality of pixel array patterns obtained in advance to find a matching pattern. Next, the coefficient of the length of the crack contributed by the central pixel of the matching pattern (predetermined) is added. Finally, the crack length in the unit area is calculated based on the total coefficient.

Therefore, as shown in FIG. 1 to FIG.
All pixel array patterns (ie, 2 ^(9-1) = 256 ways) in which the central pixel of the × 3 binary pixels is 1 are obtained in advance. However, since there are a plurality of similar pattern patterns, the number of prepared pixel array patterns can be reduced by omitting the similar pixel array patterns. 1 to 8, the numbers in parentheses (1 to 256) outside the frame of the pixel array pattern indicate the numbers of the 256 pixel array patterns.

Next, the coefficient by which the central pixel of the 3 × 3 pixel group contributes to the length of the crack (that is, 1/2 of the contour of the crack) is approximated according to the state of the surrounding 8 pixels. Numerical values in the lower part outside the frame of the pixel array patterns of FIGS. 1 to 8 indicate coefficients of the length of the crack contributed by the central pixel according to the state of eight pixels around the central pixel.

If the size of each pixel in FIGS. 1 to 8 corresponds to, for example, 2 cm × 2 cm on the road surface, the length of the crack contributed by the central pixel of the second pixel pattern has a coefficient of 1, The length of the crack contributed by the central pixel of the third pixel pattern is about 2 cm, and since the coefficient is about 1.4, it corresponds to about 2.8 cm. The same applies hereinafter.

The coefficient of the crack length contributed by the central pixel of the pixel pattern is obtained as follows. A 3 × 3 pixel group (image) centering on the central pixel is classified by approximating to the six patterns shown in FIG. It should be noted that the rotationally symmetrical and line symmetrical patterns with respect to these patterns are also the same pattern. For the patterns that do not belong to any of the six patterns in FIG. 9, the coefficient is determined in consideration of their validity. (1) First pattern ((1) in FIG. 9): In this pattern, one of the horizontal cracks (upper side) is considered to be observed, so the contribution of the central pixel to the crack length is 0. .5 pixels (that is, a coefficient of 0.5). (2) Second pattern ((2) in FIG. 9): In this pattern, it is considered that the upper and lower sides of the horizontal crack are observed, so the contribution of the central pixel to the crack length is one pixel. (That is, a coefficient of 1). (3) Third pattern ((3) in FIG. 9): In this pattern, it is considered that one side of the crack with a slope of 1 (45 degrees) is observed, so the contribution of the central pixel to the crack length is √2 / 2 pixels (that is, coefficient √2 / 2). (4) Fourth pattern ((4) in FIG. 9): In this pattern, it is considered that both sides of a crack with a slope of 1 (45 degrees) are observed, so the contribution of the central pixel to the crack length is √2 pixels (that is, coefficient √2). (5) Fifth pattern ((5) in FIG. 1): In this pattern, it is considered that one of the cracks with a slope of 1/2 is observed, so the contribution of the central pixel to the crack length is √5. / 4 pixels (that is, coefficient √5 / 4). (6) Sixth pattern ((6) in FIG. 1): In this pattern, it is considered that one of the cracks with a slope of 1/2 is observed, so the contribution of the central pixel to the crack length is √5. / 2 pixels (that is, coefficient √5 / 2).

As described above, the coefficient of the crack length contributed by the central pixel of the pixel array pattern shown in FIGS. 1 to 8 is obtained, and all the patterns and the coefficients are stored in an appropriate storage means.

Hereinafter, FIG. 1 to FIG. 8 and FIG. 10 to FIG.
A method for measuring the crack length within a unit area will be described with reference to 2.

FIG. 10 shows a method for measuring the crack length.
FIG. 20 is a diagram schematically showing how the binary image shown in FIG. 19 is divided into a plurality of pixel areas 30a to 30f each having a unit area with a center line as a reference. FIG. 11 is an enlarged view of the shaded portion 32 of the pixel region 30d of FIG. In FIG. 11, each of a plurality of squares represents one pixel, a shaded pixel (having a value of 1) represents a cracked portion, and a white pixel (having a value of 0) represents a portion not belonging to the cracking. .

Pixels at the end of the pixel area of the unit area are not surrounded by other pixels, so that the pixel P1 indicates the arrow Y1.
In the direction of, up to a pixel (not shown) inside the rightmost pixel, 0 or 1 is checked for each pixel.

As described above, when the number of pixels is 0, it is not a cracked portion and is ignored. When the number of pixels is 1, the pixel pattern determined by the state of the surrounding 8 pixels is stored in advance in an appropriate storage means. By comparing with the 256 pixel patterns shown in FIG. 1 to FIG. 8, the coefficient of the crack length contributed by the central pixel of the matched pixel pattern is calculated and added.

Next, from the pixel P2 in the third row to the pixel inside the right end pixel (not shown) along the arrow Y2, the same process as described above is performed. Further, with respect to the pixels belonging to the fourth row, the same processing is performed from the pixel P3 along the arrow Y3. In this way, if the above-mentioned processing is performed for all pixels except the outermost peripheral pixel in the pixel area of the unit area, the sum of the coefficient values of the crack length in the unit area can be obtained. FIG. 19
Since the correspondence between the length of one side of the pixel and the actual length on the road surface is known, the actual crack length in the unit area can be approximately calculated if the sum of the coefficients is calculated.

FIG. 12 is a diagram for explaining the above operation in more detail. Since pixel P1 is a pixel representing a cracked portion,
An array pattern of 3 × 3 pixels with P1 as the central pixel (see FIG.
2 (A portion surrounded by a thick line) is compared with the plurality of pixel array patterns shown in FIGS. 1 to 8. In the case of this embodiment, since it matches the pattern No. 249 in FIG.
The 46th crack length coefficient is stored in an appropriate storage location. Since the pixel P1 (1) to the right of the pixel P1 is also a pixel representing a cracked portion, the pixel array pattern surrounded by the thick line in FIG. 12B is compared with the plurality of patterns in FIGS. In this case, since it matches the 206th pattern in FIG. 6, the coefficient of the crack length assigned to the pattern is calculated and added to the coefficient already calculated. In the case of FIG. 12C, the pixel P1 (2) is not the pixel in the cracked portion, so this pixel is ignored and the processing is moved to the next pixel. Regarding (E), (F) and (I) of FIG.
Since 2 (1), P2 (2), and P3 (2) are pixels that represent cracks, respectively, the pixel patterns surrounded by thick lines are compared with the pixel array patterns of FIGS. 1 to 8, and the obtained coefficients are added. . On the other hand, in (D), (G) and (H) of FIG. 12, the pixels of interest P2, P3 and P3 (1) are not pixels in the cracked portion and are ignored.

A device (hardware) itself for carrying out the method according to the present invention is obvious to those skilled in the art, and therefore, illustration and description thereof will be omitted.

In the above description, the pixel at the end of the unit area to be measured is ignored because there is no pixel surrounding the pixel. However, when the pixel at the end portion in the unit area of the measurement target is adjacent to the pixel in the other measurement region, the adjacent pixel is used to form the array pattern of 3 × 3 pixels in the pixel array of FIGS. 1 to 8. You may make it compare with a pattern.

Further, in this embodiment, when the pixel to be processed represents a cracked portion, eight surrounding pixels (that is, 3 × 3 pixels) are extracted and compared with the pattern of FIG.
You may make it extract 5 pixels. In this case, the number of pixel array patterns prepared in advance is 2 ^(25-1) . Of course, it is the same as the case of 3 × 3 pixels that the creation and storage of the pixel array pattern can be omitted for the similar pattern.

[0037]

According to the present invention, a predetermined pixel group centered on a pixel in a cracked portion is compared with a plurality of pixel array patterns prepared in advance. By sequentially repeating this comparison operation for the pixels in the cracked portion, the crack length in the unit area can be obtained. In particular, since the method according to the present invention can be carried out with extremely simple software, it has a remarkable effect that the crack length of the road surface can be measured in an extremely short time as compared with the conventional method of tracing the crack as a line.

[Brief description of drawings]

FIG. 1 is a diagram showing a pixel array pattern of 3 × 3 pixels necessary for implementing the present invention.

FIG. 2 is a diagram showing a pixel array pattern of 3 × 3 pixels necessary for implementing the present invention.

FIG. 3 is a diagram showing a pixel array pattern of 3 × 3 pixels necessary for implementing the present invention.

FIG. 4 is a diagram showing a pixel array pattern of 3 × 3 pixels necessary for implementing the present invention.

FIG. 5 is a diagram showing a pixel array pattern of 3 × 3 pixels necessary for implementing the present invention.

FIG. 6 is a diagram showing a pixel array pattern of 3 × 3 pixels necessary for implementing the present invention.

FIG. 7 is a diagram showing a pixel array pattern of 3 × 3 pixels necessary for implementing the present invention.

FIG. 8 is a diagram showing a pixel array pattern of 3 × 3 pixels necessary for implementing the present invention.

FIG. 9 is a diagram showing a basic pattern for obtaining a coefficient of a crack length contributed by a central pixel of a pixel array pattern of the present invention.

FIG. 10 is a diagram showing how a binary image is divided into a plurality of “unit area pixel regions” in order to explain an embodiment of the present invention.

FIG. 11 is a diagram illustrating details of a hatched portion of a pixel area 30d having the unit area of FIG. 10 to explain an embodiment of the present invention.

FIG. 12 is a diagram for explaining an example of the present invention.

FIG. 13 is a digital image in a digital image processing process for obtaining a digital image used in the present invention.

FIG. 14 is a digital image in a digital image processing process for obtaining a digital image used in the present invention.

FIG. 15 is a digital image in a digital image processing process for obtaining a digital image used in the present invention.

FIG. 16 is a digital image in a digital image processing process for obtaining a digital image used in the present invention.

FIG. 17 is a digital image in a digital image processing process for obtaining a digital image used in the present invention.

FIG. 18 is a digital image in a digital image processing process for obtaining a digital image used in the present invention.

FIG. 19 is a digital image used in the present invention.

[Explanation of symbols]

 30a unit area pixel area 30b unit area pixel area 30c unit area pixel area 30d unit area pixel area 30e unit area pixel area 30f unit area pixel area P1 pixel P2 pixel P3 pixel

Claims (4)

[Claims] 1. A road surface is photographed to obtain an analog image of the road surface, image processing is performed on the analog image to obtain a binary image representing presence or absence of cracks, and the crack length per unit area of the road surface is obtained from the binary image. (A) Binary N × N pixel (N) in which the central pixel represents a crack (N)
A plurality of pixel array patterns that can be taken by an odd number of 3 or more) are obtained in advance, and a coefficient of the length of the crack contributed by the central pixel representing the crack is previously obtained for each of the plurality of pixel array patterns, ) A unit area is set in the binary image representing the crack, and (c) N × N pixels centered on the pixel representing the crack are extracted from pixels in the unit area, and (d) the extracted N × N. The pixel array pattern of the pixels is compared with the previously determined plurality of pixel array patterns to obtain the same pixel array pattern, and the coefficient of the crack length given to the pixel array pattern is obtained. (C) to (d) are repeated for each pixel in the unit area, the coefficients obtained respectively are added, and the crack length in the unit area is measured based on the total of the obtained coefficients. Do Method of measuring the crack lengths surface. 2. The method for measuring a crack length of a road surface according to claim 1, wherein N is 3. 3. The plurality of pixel array patterns is 2
^The method for measuring a crack length of a road surface according to claim 1, wherein the number is ^{(N · N-1)} . 4. The plurality of pixel array patterns is 2
^The method for measuring a crack length of a road surface according to claim 1, wherein the method comprises a pixel array pattern obtained by removing a similar pixel pattern from ^{(N · N-1)} pixel array patterns.