JP2008139285A - Construct using picture processing technique, and crack width measuring method of product - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To enhance a picture data processing speed and to enhance a resolution of an inspection system, and cutting down construction cost of crack inspection system. <P>SOLUTION: After carrying out a monochrome picture processing of an original picture of a construct obtained through a camera or a product, several sorts of filtering operations are performed to measure the crack width by a crack analysis in order to extract only a crack. The crack analysis contains the process of measuring the crack width with a concentration value histogram of a pixel distributed in the cracks, using the concentration value in the crack part. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a method capable of measuring the crack width according to the density value of a pixel in surface crack measurement of structures and products using image processing techniques, and more particularly, searching for a surface crack from an image received through a camera. When measuring the width, low resolution video processing is realized by implementing a new method that measures the crack width based on the density value of the pixel that is differentiated from the method that uses the number of pixels that has been used a lot. The present invention relates to a method for measuring a crack width of a structure and a product using a video processing technique that can be measured by a technique and can perform a high-speed processing in terms of data processing speed.

  In general, concrete structures such as tunnels may cause peeling or collapse of water pressure, earth pressure, earthquakes, structural defects, etc. over time.

  For these reasons, safety inspections and precise safety diagnosis must be carried out regularly, and the state and stability of concrete structures are evaluated, and stability is ensured by repairing, reinforcing, and restoring performance. It must be maintained.

  The most basic item is the appearance survey.

  In particular, visual inspection for concrete cracks is an important factor in evaluating the internal and external conditions of structures and determining the performance items and methods of local precision safety diagnosis.

  However, if a structure such as a tunnel with a large cross section and a long extension length is investigated by an existing visual inspection method using the naked eye, it takes a long time and it is difficult to perform an objective investigation.

  Therefore, it is necessary to automate the survey method and improve the accuracy and speed. In recent years, an appearance survey technique and system using a laser, a CCD camera, etc. have been developed and used gradually.

  Each system has advantages and disadvantages, but among them, for example, a method for investigating the appearance state through an image obtained by using a CCD camera as described in Non-Patent Document 1 has a high performance with respect to construction cost. It is a favored trend because it is superior to other systems due to advantages such as survey time.

  A system for detecting cracks in a tunnel using a CCD camera and digital image processing has been widely used both in Japan and overseas, and in Japan, in particular, it is often used as an appearance inspection method for tunnel safety diagnosis.

  In addition, a portable system that easily detects cracks using a CCD camera, a tunnel crack detection system using a single CCD camera, a CCD camera such as a long-distance structure crack detection device using a CCD camera, etc. Continuous crack detection systems are being developed.

  FIG. 1 shows the principle of the crack detection system 10 using the digital video processing technique.

  This is a method in which an image of a concrete structure is acquired and analyzed using the CCD camera 11, and a crack portion on the surface is detected to obtain information values such as width and length.

  By using this method, it is possible to conduct a quicker and more accurate investigation than existing human visual examination methods.

  Especially in the case of structures such as tunnels (subways, road tunnels, waterway tunnels, etc.) that have a long extension and a large cross section, the use of this method has a great effect in terms of time, economy, and accuracy. Is obtained.

  As shown in FIG. 1, a video signal obtained by using a CCD camera 11 from the surface of a concrete structure is converted into a digital signal through a digitizer 12 and stored as image data 13. The analysis information (width, length, etc.) 15 and the crack processing image 16 for the crack are obtained. The data stored in the image 13 is displayed on the display 17.

  Factors affecting here include the resolution of the CCD camera 11, detection algorithm, illumination conditions, vibration, movement, speed, lens distortion, angle between the CCD camera 11 and the imaging surface, and the like.

  In particular, the resolution and detection algorithm of the CCD camera 11 have the greatest influence on the error of the crack measurement value.

  First, looking at the influence of the resolution of the CCD camera 11, in order to calculate the width, length, etc. of the crack, it is necessary to know the actual length represented by one pixel. It can be obtained if the surface area of the concrete and the number of pixels of the CCD camera 11 are known.

Pl = W _area / R _CCD

Here, Pl is the actual length (mm / pixel) represented by one pixel, W _area is the length of one side of the imaging surface (eg, the length of the Y axis), and R _CCD is W _area and This represents the number of pixels of the CCD camera 11 in the same direction (eg, the number of pixels on the Y axis).

  The minimum width of crack measurement is determined in consideration of the pixel size of the above formula when photographing a concrete structure.

  That is, in the case of the crack width detection method based on the calculation of the number of pixels, which is a general crack detection method, when attempting to measure up to a crack width of 0.1 mm, the actual length of one minimum pixel is 0.1 mm or less. Must be.

  Therefore, the area to be photographed is set by the angle of view of the CCD camera 11 in consideration of the minimum width of crack measurement.

  However, the accuracy of the width calculation using the number of pixels decreases as the number of pixels in the crack width decreases.

  This is because the pixel has a quadrangular shape, and therefore, when the crack is not in the vertical and horizontal direction, a difference occurs in the actual length of one pixel.

  When utilizing a system that currently uses such a method, sufficient consideration must be given to such problems.

  Next, looking at the influence of the detection algorithm, the detection algorithm is operated as follows.

  The video signal that has entered through the CCD camera 11 is converted into a digital signal through the digitizer 12 and stored in an image storage 13.

  The stored data goes through a known video processing stage for crack detection and measurement.

  For example, as shown in FIGS. 2 and 3, after performing black and white image processing 22 of the original image 21 obtained through the CCD camera 11, several kinds of filtering operations are performed in order to extract only cracks by the crack detection filter 23. After that, response merging 24, crack fragment detection 25, and crack analysis 26 are performed. Finally, the width, length and coordinates 28 are calculated by the number of pixels in the crack through the data analysis 27.

Here, the original image 21 is, for example, a 24-bit or 32-bit color image that has entered through the CCD camera 11. In the black-and-white video processing 22, the color information of the color image having the information amount of the left bit number (scale) is removed and converted to a gray image having the information amount of the bit number (gray scale) smaller than the bit number. That is, color information is unnecessary when extracting information such as crack width and length in crack detection and measurement.
For example, in the monochrome image processing 22, a color image having a scale of 24 bits or 32 bits is converted into a gray image having a gray scale of 256 gradations (8 bits) from 0 to 255. By performing this conversion, the file capacity in the video processing can be greatly reduced.

  In the crack detection filter 23 process, various image processes are performed in the black and white video process 22 in order to detect crack information from an image converted into a gray scale. For example, in order to extract only cracks (cracks) as the primary processing, the image converted to gray scale in the black and white video processing 22 is subjected to blank processing below the density of cracks and primary filtering is performed.

In the response merging 24 performed after the crack detection filter 23 processing, a video image subjected to primary filtering by the crack detection filter 23 processing is obtained.
In the crack fragment detection 25, the crack information in the video image information that is primarily completed through the response merge 24 is displayed on the display 17 so that it can be confirmed whether cracks or non-cracks are possible.
The crack analysis 26 refers to a crack measurement stage, in which the width, length, coordinates, etc. of the crack are measured, and these information can be finally output (28). After the crack analysis 26, the data analysis 27 is performed by calculating the width, length, and coordinates of the crack using the information of all pixels.

  In the process of obtaining the crack width, conventionally, the width is measured by obtaining the number of image pixels in the crack portion. At this time, the resolution of the crack width measurement depends on the resolution of the video apparatus.

  Therefore, in order to measure a crack width of 1 mm or less, the angle of view must be narrowed or the resolution must be increased, and an expensive imaging device that can capture images with a small inspection area or obtain high resolution must be used. Therefore, the processing time increases as the data capacity increases.

  For example, in order to measure a crack width of 0.1 mm over an inspection area of 1 m × 1 m, a minimum of 100 million pixels of 10000 pixels × 10000 pixels is required, so that one side of one pixel occupies 0.1 mm In order to actually improve the accuracy, more pixels are required.

  Therefore, even if several video devices are used in combination to increase the resolution, it is impossible to measure the crack width of large structures and large products by conventional methods in terms of system configuration costs, data processing and operation. is there.

“Crack extraction method of concrete surface by image processing”, Keisuke Fujita et al. Nondestructive Inspection Vol.56 No.7 (2007) p. 371-p. 377

  Accordingly, in the crack width measurement of structures and products using image processing techniques, it is desired to reduce the construction cost of a crack inspection system, improve the image data processing speed, and increase the resolution of the inspection system.

  The method for measuring the crack width of structures and products using the image processing technique of the present invention (hereinafter sometimes simply referred to as “the crack width measuring method of the present invention”) is a method for processing a black and white image of an original image obtained through a camera. In the method of measuring the crack width of a structure and product using a video processing technique for performing several kinds of filtering operations to extract only cracks and measuring the crack width through crack analysis after performing the crack analysis, The method includes a step of measuring the width of a crack with a histogram of density values of pixels distributed in the crack.

  The process of measuring the width of the crack with the histogram of the density values of the pixels distributed in the crack is performed after a calibration bar having a plurality of grooves having different widths is placed on the surface of the inspection target having a crack and photographed together. The method is characterized in that the density value histogram of the groove distributed on the calibration bar is compared with the density value histogram of the actual crack pixel.

  Further, the calibration bar is characterized by using a groove having the groove formed with a width of 0.1 mm to 0.1 mm.

  According to the present invention, the crack width is measured using the pixel density value histogram in the crack measurement using the image processing, and therefore, a higher resolution than the conventional crack width measurement method using the number of pixels is not required. The data capacity can be reduced, and the data processing speed can be increased and crack inspection can be performed quickly and accurately.

  Hereinafter, the crack width measuring method of the present invention will be described in detail according to embodiments of the present invention with reference to the accompanying drawings.

  The crack width measuring method of the present invention basically includes a digital image processing process of a crack inspection system.

That is, the process of performing black and white image processing 22 on the original image 21 photographed by the camera shown in FIG. 2, the process of filtering for crack detection by the crack detection filter 23, and the response merge 24 performed after the process of the crack detection filter 23, This includes a process of obtaining a primary filtered video image, a process of detecting (25) crack fragments, a process of measuring crack width 28 through crack analysis 26 and data analysis 27, and the like.
The basics of the processes indicated by reference numerals 22 to 28 are the same as those of the conventional digital video processing process shown in the block diagram of FIG. 2 except for the point of the present invention described below.

  As can be seen from FIG. 3, the crack 31 composed of square pixels finely arranged on the crack surface can be obtained from the above-described image processing process.

  Normally, the number of pixels in the crack 31 is calculated here and the crack width is measured, but the embodiment of the present invention provides the following new method.

  FIG. 4 illustrates an example representing pixels distributed in the cracks 41 of the original image 21 (see FIG. 2).

  As shown in FIG. 4, it can be seen that the pixels are distributed in the crack width corresponding to 0.1 to 1 mm for the entire crack 41.

  When shooting is performed with settings as described above, the video data appears as shown in FIG.

As an operation after photographing a cracked surface using a camera, in order to represent only the crack of the photographed original image 21 (the density value of the crack always has a lower value than the background), the calibration of FIG. A value exceeding the density value of the groove 61 having a width of 0.1 mm corresponding to the smallest width of the motion bar 60 is taken out from the density value histogram of the entire image surface of the original image 21, and the remaining non-cracked portion (background) If the density value 255 is white, a crack 51 as shown in FIG. 5 is represented.
Here, the density value histogram is a display showing density values. For example, the density value histogram is a color in a black-and-white image or a color image defined according to the density value. In this case, the color of the density value histogram and the width of the crack (crack) correspond to each other via the density value, whereby the width of the crack (crack) is obtained based on the density value histogram.

  As illustrated in FIG. 5, it can be seen that the pixels distributed in the crack 51 have different density value histograms.

  In the embodiment of the present invention, the width of the crack is measured by the density value histogram at this time.

  For this purpose, the crack width of the density value histogram of the pixel is measured using the crack calibration bar 60.

  As shown in FIG. 6, the crack calibration bar 60 is made of a steel material having a thickness of 10 mm or more, and the grooves 61 to 70 having a width of 0.1 mm up to 0.1 to 1 mm have an error range of ± 0. It consists of a bar shape processed at 0.01 mm. Here, the crack calibration bar 60 illustrated in FIG. 6 is photographed together with the structure having the crack 31 under the same photographing condition as the photographing condition of the structure surface having the crack 31.

  When the crack calibration bar 60 is photographed and then image processing is performed, density value histograms corresponding to the grooves 61 to 70 can be obtained.

  For example, as illustrated in FIG. 7 and FIGS. 8A to 8I, density value histograms of different colors of grooves corresponding to 0.1 to 0.9 mm can be obtained.

  FIG. 8A shows the crack width processing result by the image processing of the crack 0.1 mm, that is, the color of the density histogram, FIG. 8B is 0.2 mm, and FIG. 8C is 0.3 mm. 8 (d) is 0.4 mm, FIG. 8 (e) is 0.5 mm, FIG. 8 (f) is 0.6 mm, FIG. 8 (g) is 0.7 mm, and FIG. 8 (h) is 0.8 mm. 8 (i) represents 0.9 mm.

  Of course, here, the image is shown in which the grooves 61 to 70 of the crack calibration bar 60 are not photographed directly, and the images are photographed while adjusting the grooves between the two blocks at intervals of 0.1 mm. The same processing results as those shown in FIGS. 8A to 8I can be obtained also when video processing is performed after shooting the grooves 61 to 70 of the motion bar 60.

  Therefore, after such a crack calibration bar 60 is placed on the surface of the inspection target having a crack and photographed together, the density value histogram of the grooves 61 to 70 distributed in the calibration bar 60 and the pixel over the actual crack width are displayed. The crack width can be measured by comparing the density value histograms.

  For example, if the density value histogram of the pixel in the crack matches the density value histogram of the groove 81 shown in FIG. 8A, the crack width at this time can be calculated as 0.1 mm.

  That is, when an attempt is made to extract a crack width of 0.1 mm or more as the minimum measurement value, the actually captured image is a density value corresponding to 0.1 mm of the crack calibration bar 60 shown in FIG. Using a)), the above calculation can be performed by performing a process of adjusting a pixel having a density value equal to or lower than that value in an actual captured image to a density value of zero.

  Here, after comparing the density value histogram obtained from the crack calibration bar 60 and the density value histogram obtained from the actual crack surface, a method for obtaining a crack width by converting the value into a numerical value through a crack measurement algorithm is described in the art. Any generally known method can be adopted without any particular limitation.

  However, the present embodiment is different from the prior art in that all pixels of the original image 21 photographed are used in the prior art, whereas in this embodiment, the density of each crack distributed in the crack calibration bar 60 is used. The value is compared with the density value of the actual crack pixel.

  In addition, the method of quantifying this through the crack measurement algorithm and obtaining the crack width is the area labeling process for grouping by the connection characteristics of density value areas extracted from the video, the primary filtering process, the pixel characteristics, the noise There are filtering techniques such as the process of connecting adjacent areas to the crack cutting phenomenon caused by, and the calculation of the feature shape for classification of cracks and non-cracks (white spots, shadows, stains, etc.).

  Table 1 below shows a conventional method for measuring crack width using the number of pixels and an embodiment of the crack width measuring method of the present invention (embodiment of the present invention).

  As can be seen from Table 1, in the embodiment of the present invention, the crack width can be measured even in a low-resolution video device by measuring the crack width using the density value histogram from 0 to 255 of the pixel. In addition, it is possible to perform fast processing in terms of data processing speed.

It is the schematic showing the structure of the crack inspection system using the conventional video processing technique. It is a block diagram showing a digital video processing process in a crack inspection system using a conventional video processing technique. It is a photograph showing a video processing and crack analysis state in a crack inspection system using a conventional video processing technique. It is a photograph which shows the pixel distributed in the crack at the time of imaging | photography in embodiment of this invention. In an embodiment of the present invention, it is a photograph showing the characteristic of a pixel in which a density value histogram is expressed differently. It is a front view showing the crack calibration bar used in the embodiment of the present invention. It is a photograph showing the density value histogram of the crack width used in embodiment of this invention. It is a photograph showing the density value histogram of the crack width of 0.1 mm in embodiment of this invention. It is a photograph showing the density value histogram of crack width 0.2mm in embodiment of this invention. It is a photograph showing the density value histogram of the crack width of 0.3 mm in embodiment of this invention. It is a photograph showing the density value histogram of the crack width of 0.4 mm in embodiment of this invention. It is a photograph showing the density value histogram of the crack width of 0.5 mm in embodiment of this invention. It is a photograph showing the density value histogram of the crack width 0.6mm in embodiment of this invention. It is a photograph showing the density value histogram of the crack width 0.7mm in embodiment of this invention. It is a photograph showing the density value histogram of crack width 0.8mm in embodiment of this invention. It is a photograph showing the density value histogram of crack width 0.9mm in embodiment of this invention.

Explanation of symbols

21 ... Original image, 22 ... Monochrome image processing, 23 ... Crack detection filter 26 ... Crack analysis, 31 ... Crack (conventional method)
41 ... crack (original image), 51 ... crack (crack detection image)
60 ... Calibration bar, 61-70 ... Groove

Claims (3)

After processing the original image obtained through the camera in black and white, several types of filtering work is performed to extract only cracks, and the crack width of structures and products is measured using image processing techniques that measure crack width through crack analysis. In the method
A method of measuring a crack width of a structure and a product using an image processing technique, comprising: a step of measuring a crack width using a density value histogram of pixels distributed in the crack in the crack analysis.   In the process of measuring the width of the crack using the histogram of density values of pixels distributed in the crack, a calibration bar (Calibration Bar) having a plurality of grooves having different widths is arranged on the surface of the inspection target having a crack and taken together. 2. The image processing technique according to claim 1, wherein a density value histogram of the groove distributed in the calibration bar is compared with a density value histogram of an actual crack pixel. Crack width measurement method for structures and products.   The image processing technique according to claim 1 or 2, wherein the calibration bar has the groove formed with a width of 0.1 mm to 0.1 mm. Method for measuring the crack width of manufactured structures and products.