Image processing system, image processing method and program

Abstract

To provide an image processing system that accurately detects crack information in newly captured new image data based on old image data. [Solution] The system is characterized by comprising an input unit that inputs old image data and new image data depicting cracks on the wall surface to be inspected, a crack feature recognition unit that recognizes the shape features and positions of old cracks depicted in the old image data, a pre-processing unit that recognizes lines and dotted lines in the new image data and thickens them to make adjacent lines and dotted lines into a series of thick lines, a crack detection unit that, after processing by the pre-processing unit, detects areas containing old cracks based on the shape features and position information of the cracks recognized by the crack feature recognition unit and shows them in the new image data so that they can be distinguished from other lines, and an output unit that outputs the image data after processing by the crack detection unit. By adopting this configuration, accurate crack detection processing is achieved.

Description

[Technical Field]

[0001] The present invention relates to a technology for checking the condition of cracks on wall surfaces, and more particularly to an image processing system, an image processing method, and a program used therefor that can be used, for example, to periodically check the condition of cracks on the inner wall surfaces of tunnels. [Background technology]

[0002] In the past, the applicant proposed an image processing system used to check the status of cracks on the wall surface to be inspected, which processes image data obtained by photographing the inner wall surface of a tunnel and converts it into a plan view to make it easier to verify (Patent Document 1).The plan view on which information about the cracks is entered is called a deformation development view. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 2017-201465 Summary of the Invention [Problem to be solved by the invention]

[0004] For example, the accuracy of the crack position and shape shown in a deformation development map in which crack information is hand-drawn on a floor plan (hereinafter referred to as a hand-drawn deformation development map) depends on the skill level of the worker. When comparing a hand-drawn deformation development map with a deformation development map based on image data obtained by subsequent photography, there are errors (inaccuracies) in the accuracy of the fine crack shapes. Furthermore, in image data obtained by photography, cracks that previously appeared as a line may appear as a dotted line due to differences in photography conditions (photographic equipment, photography position, lighting, sunlight conditions, etc.).

[0005] For the reasons stated above, when an image processing device is used to check information about cracks that appear in new image data obtained by photographing based on old image data obtained by scanning a hand-drawn deformation development diagram, or when checking information about cracks that appear in new image data obtained by photographing based on old image data, even though the photograph was taken, accurate results may not be obtained due to errors caused by the precision of the hand-drawing or the conditions of the photograph.

[0006] The present invention has been made to solve the above-mentioned conventional problems, and aims to provide an image processing system, an image processing method, and a program for use in said system or method, which allows for easy and accurate confirmation of information about cracks appearing in newly captured image data (new image data) based on information about cracks described in previous deformation development drawings (old image data), making it easier to visually recognize the progression of cracks, etc. [Means for solving the problem]

[0007] In order to achieve the above object, the image processing system of the present invention includes an input unit that inputs old image data depicting old cracks on the inspection wall surface and new image data depicting cracks on the inspection wall surface at a time newer than the old image data, and The binary image data obtained by edge detection processing A crack feature recognition unit that recognizes the shape features and position of the drawn old crack based on the new image data. The binary image data obtained by edge detection processing Recognizes lines and dotted lines, and thickens the recognized lines and dotted lines By adding a border to the edge of the binarized image, The system is characterized by comprising a pre-processing unit that converts adjacent lines and dotted lines into a series of thick lines, a crack detection unit that detects locations containing old cracks from the series of thick lines obtained by processing by the pre-processing unit based on information on the shape characteristics and position of the old crack (which has not been subjected to thickening processing) recognized by the crack feature recognition unit, and displays the series of thick lines containing the old crack in the new image data so that they can be distinguished from other lines, and an output unit that outputs, displays, or prints out the image data after processing by the crack detection unit.

[0008] Furthermore, it is preferable that the crack detection unit indicates in the new image data a series of thick lines including the old crack so that the old crack location can be distinguished from other locations.

[0009] It is preferable that the output unit is equipped with a display monitor and a pointing device, and when multiple locations are detected in the series of thick lines obtained by processing by the pre-processing unit that are judged to include the old crack due to errors in the information on the shape characteristics and position of the old crack, the crack detection unit displays the multiple detected locations on the display monitor, and indicates the correct location specified by the pointing device as the location containing the old crack, and shows the series of thick lines including the old crack in the new image data so that they can be distinguished from other lines.

[0010] The old image data is preferably image data obtained by scanning a plan view on which old cracks on the inspected wall surface are hand-drawn.

[0011] Furthermore, it is preferable that the crack detection unit is provided with a new crack confirmation unit that identifies newly formed cracks and deletes lines that are not cracks for other lines that are determined not to be a series of thick lines including old cracks.

[0012] The image processing method for solving the above problem includes an input step of inputting old image data depicting old cracks on the inspection wall surface and new image data depicting cracks on the inspection wall surface at a time newer than the old image data, and The binary image data obtained by edge detection processing A crack feature recognition process for recognizing the shape features and position of the drawn old crack based on the new image data. The binary image data obtained by edge detection processing Recognizes lines and dotted lines, and thickens the recognized lines and dotted lines By adding a border to the edge of the binarized image,The method is characterized by comprising a preprocessing step of converting adjacent lines and dotted lines into a series of thick lines, a crack detection step of detecting locations containing old cracks from the series of thick lines obtained by the preprocessing step based on information on the shape characteristics and position of the old cracks recognized in the crack feature recognition step, and displaying the series of thick lines containing the old cracks in the new image data so that they are distinguishable from other lines, and an output step of outputting, displaying, or printing out the image data processed by the crack detection step.

[0013] Furthermore, a computer-readable image processing program for solving the above problem includes an input unit that inputs old image data depicting old cracks on the inspection wall surface and new image data depicting cracks on the inspection wall surface at a time newer than the old image data into a computer by executing the image processing program; and The binary image data obtained by edge detection processing A crack feature recognition unit that recognizes the shape features and position of the drawn old crack based on the new image data. The binary image data obtained by edge detection processing Recognizes lines and dotted lines, and thickens the recognized lines and dotted lines By adding a border to the edge of the binarized image, The system is characterized by comprising a pre-processing unit that converts adjacent lines and dotted lines into a series of thick lines, a crack detection unit that detects locations containing old cracks from the series of thick lines obtained by processing by the pre-processing unit based on information on the shape characteristics and position of the old crack recognized by the crack feature recognition unit, and displays the series of thick lines containing the old crack in the new image data so that they can be distinguished from other lines, and an output unit that outputs, displays, or prints out the image data after processing by the crack detection unit. [Effects of the Invention]

[0014] According to the present invention, the lines and dotted lines appearing in the new image data are thickened, The edges of the binarized image are bordered to turn adjacent lines and dotted lines into a series of thick lines, In order to detect the locations containing old cracks that appear in the old image data, errors caused by the accuracy of hand-drawing or the conditions of photography may occur. Edge detection processing can detect lines that were originally continuous but have become discontinuous after binarization. To reduce false detections of no relevant cracks and realize rapid crack detection processing. [Brief explanation of the drawings]

[0015] [Figure 1] 1 is a configuration diagram of an image processing system according to an embodiment of the present invention. [Figure 2] FIG. 2 is a functional block diagram of the image processing system. [Figure 3] Main Routine [Figure 4] 10 is a flowchart of a process for reading old image data. [Figure 5] A deformation development diagram showing crack information on a plan view based on old image data. [Figure 6] 10 is a flowchart of a process for recognizing shape characteristics and positions of cracks. [Figure 7] 10 is a flowchart of a new image data reading process. [Figure 8] Plan view based on new image data. [Figure 9] 10 is a flowchart of preprocessing. [Figure 10] A portion of the new image data. [Figure 11] The result of edge detection processing on the new image data. [Figure 12] Image data after fattening. [Figure 13] 10 is a flowchart of a crack detection process. [Figure 14] Flowchart for identifying new cracks. [Figure 15] The old cracks drawn in the old image data are shown with thin lines, cracks extending from the old cracks with thick lines, and potentially new cracks with hollow lines. DETAILED DESCRIPTION OF THE INVENTION

[0016] The image processing system according to the present invention includes an image data input unit, a crack feature recognition unit, a preprocessing unit, a crack detection unit, and an output unit. The input unit inputs old image data depicting old cracks on the wall surface to be inspected and new image data depicting cracks on the wall surface at a time more recent than the old image data. Specifically, the input unit inputs old image data obtained by scanning a development diagram of the wall surface to be inspected (hereinafter referred to as a hand-drawn development diagram of the wall surface) in which crack information (old cracks) has been hand-drawn on the floor plan of the wall surface to be inspected, or image data obtained by photographing the old image data with an older photographed date and new image data with a more recent photographed date and time. The crack feature recognition unit recognizes the shape features and position of the depicted old cracks based on the old image data. The preprocessing unit recognizes lines and dotted lines based on the new image data and thickens the recognized lines and dotted lines to make adjacent lines and dotted lines into a series of thick lines. The crack detection unit detects locations containing old cracks from the series of thick lines obtained by processing in the preprocessing unit based on the information on the shape features and position of the old cracks recognized by the crack feature recognition unit, and displays the series of thick lines containing the cracks in the new image data so that they can be distinguished from other lines. The output unit outputs, displays, or prints out the image data processed by the crack detection unit. By adopting this configuration, it is possible to reduce false detections of no cracks due to errors caused by hand-drawing accuracy or photography conditions, thereby achieving rapid crack detection processing.

[0017] An image processing system 1 according to an embodiment of the present invention will be described with reference to the accompanying drawings. FIG. 1 shows a configuration diagram of the image processing system 1. The system 1 comprises a central processing unit (CPU) 2, a ROM 3, a RAM 4, a hard disk (hereinafter referred to as HDD) 5, a scanner 6, a camera 7, a printer 8, a display monitor 9, and a pointing device 10. The ROM 3 stores an operating system program (OS). The RAM 4 is used as a work memory. The HDD 5 stores image processing programs, new and old image data, and the output of processed image data. The scanner 6 reads image data from a floor plan. The camera 7 digitally captures image data of the wall surface to be inspected. The printer 9 prints out the processed image data. The display monitor 9 displays image data during and after processing. The pointing device 10 is composed of a keyboard and mouse, a touchpad, or the like, for selecting a crack displayed on the display monitor 9 and placing it in the correct position, or for specifying the result of determining whether a line or dotted line appearing in the image data is a new crack or not.

[0018] When the system 1 starts up, the OS stored in the ROM 3 is started, and then a computer-readable image processing program is read from the HDD 5 to the RAM 4 and executed, performing a series of main routine processes described below. Execution of the image processing program realizes the image processing system 1 that performs a predetermined image processing method. Alternatively, a program storage method (von Neumann architecture) may be employed in which the OS is stored in the HDD 5 and is read from the HDD 5 to the RAM 4 and executed when the system 1 starts up.

[0019] FIG. 2 is an explanatory diagram of functional blocks realized by the image processing system 1. The functional blocks are composed of an image data input unit 21, a crack feature recognition unit 22, a preprocessing unit 23, a crack detection unit 24, a new crack confirmation unit 25, and an output unit 26. The input unit 21 reads a deformation development diagram (hereinafter referred to as a hand-drawn deformation development diagram) in which information about old cracks is hand-drawn on a plan view of the inspected wall surface, using a scanner 6, and inputs the old image data stored on the HDD 5, or old image data stored on the HDD 5 with an older photographing date and time, to the crack feature recognition unit 22. The input unit 21 also inputs new image data obtained by photographing with the camera 7 and with a newer photographing date and time than the old image data, or new image data stored on the HDD 5, to the preprocessing unit 23. The crack feature recognition unit 22 recognizes the shape features and position of the depicted old crack based on the old image data from the input unit 21. The preprocessing unit 23 recognizes lines and dotted lines based on the new image data from the input unit 21, thickens the recognized lines and dotted lines, and converts adjacent lines and dotted lines into a series of thick lines. The crack detection unit 24 detects cracks from the series of thick lines obtained by the preprocessing unit 23 based on the shape characteristics and position information of the cracks recognized by the crack feature recognition unit 22, and displays the series of thick lines including the cracks in the new image data so that they can be distinguished from other lines. The new crack confirmation unit 25 uses the display monitor 9 and pointing device 10 to identify new cracks and delete lines determined not to be cracks for other lines determined not to be cracks by the crack detection unit 24. The output unit 26 outputs the processed image data to the HDD 5, displays it on the display monitor 9, or prints it out to the printer 8. Output by the output unit 26 includes output to a terminal (not shown) connected via a network, such as a smartphone or laptop equipped with a device equivalent to a HDD and display monitor, or a mobile printer with printing capabilities. The new crack confirmation unit 25 may be omitted depending on the required system specifications.

[0020] The processes executed by the image processing program read by the CPU 2 from the HDD 5 will be described in order using flowcharts. FIG. 3 shows the main routine. First, the old image data is read (step S1), and the shape characteristics and position of the old crack are confirmed for the read old image data (step S2). Next, the new image data is read (step S3), followed by preprocessing (step S4), in which lines and dotted lines are detected for the read new image data and thickened to create a series of thick lines. Preprocessing reduces false positives, such as errors in the accuracy of the hand-drawn old crack or errors caused by the conditions under which the new image data was captured, resulting in no corresponding old crack in the new image data. In the crack detection process (step S5), locations containing old cracks within the lines and dotted lines obtained after preprocessing are detected. In the new crack confirmation process (step S6), lines and dotted lines other than the lines and dotted lines in the new image data that were detected as containing old cracks that appeared in the old image data are identified as new cracks and non-crack lines are deleted. In the output process (step S7), the old cracks that appeared in the old image data and the newly formed cracks are displayed in a distinguishable manner, and the process returns to step S1 (or the process ends).

[0021] 4 shows a flowchart (subroutine flowchart) of the old image data reading process (step S1). If the old deformation development drawing of the inspected surface has been read by the scanner 6 (Yes in step S11), the handwritten image data is read and expanded in RAM 4 (step S12), stored in HDD 5 (step S13), and the process returns to the main routine (step S14). If no image has been read by the scanner 6 (No in step S11), the old image data is read from HDD 5 and expanded in RAM 4 (step S14), and the process returns to the main routine (step S14).

[0022] FIG. 5 shows a plan view 102 obtained by photographing the inside of a tunnel 100 with a camera 7 moving in the direction of arrow 101. Lines 103 that appear as a grid in the drawing are traces of formwork used when pouring concrete into the tunnel wall. Cracks 104 are shown drawn with solid lines. When explaining how to depict cracks in the following process, we will explain the cracks in the area surrounded by a frame 105.

[0023] 6 shows a flowchart (subroutine flowchart) of the process (step S2) for recognizing the characteristic shape and position of the old crack. Edge detection is performed on the old image data stored in RAM 4 to detect lines and dotted lines in the image (step S21). The detected series of lines and dotted lines are converted into vector data to form coherent data (step S22), which is then stored in HDD 5 (step S23), and the process returns to the main routine (step S24).

[0024] 7 shows a flowchart (subroutine flowchart) of the new image data reading process (step S3). If image data of the inspection surface has been captured by the camera 7 (Yes in step S31), the image data is received from the camera 7, expanded in RAM 4 (step S32), and stored in HDD 5 (step S33), and the process returns to the main routine (step S34). If there is no captured data by the camera 7 (No in step S31), the new image data is read from HDD 5 and expanded in RAM 4 (step S34), and the process returns to the main routine (step S35). The new image data to be read from HDD 5 is selected using the display monitor 9 and pointing device 10 in a known procedure.

[0025] FIG. 8 shows a plan view 112 based on new image data obtained by photographing the inside of tunnel 100 with camera 7 moving in the direction of arrow 101. Lines 103 that appear as a grid in the drawing are traces of formwork used when pouring concrete into the tunnel wall. Cracks 114 are shown drawn with solid lines. When explaining how to depict cracks in the following process, we will explain the cracks in the area surrounded by frame 115 (the same size as frame 105).

[0026] FIG. 9 shows a flowchart (subroutine flowchart) of the preprocessing (step S4). Edge detection processing is performed on the new image data (multi-level gradation data) to detect lines and dotted lines (step S41). As a thickening process, the lines and dotted lines detected in step S41 are thickened, for example, by dilation (step S42). The lines and dotted lines after the thickening process are recognized (step S43), and the process returns to the main routine (step S44). The thickening process refers to a process of thickening lines. In addition to the dilation process, which thickens lines by increasing the number of pixels inside lines and dots, it may also be a process of thickening lines by applying a border of 3 to 5 pixels to the edge portions of the binarized image.

[0027] Fig. 10 is an enlarged view of the area surrounded by a frame 115 in the plan view 112 shown in Fig. 8. The image data is photographic image data, and is multi-level gradation data.

[0028] Figure 11 shows the binary data obtained by performing edge detection processing on the multi-level gradation data shown in Figure 10. The areas indicated by circles 116, 117, 118, and 119 were continuous in the original multi-level gradation data, but after binarization they have become discontinuous dotted lines.

[0029] FIG. 12 shows binary data obtained by thickening the lines and dotted lines shown in FIG. 11, for example, by increasing the line width by 3 to 5 pixels. It can be seen that the areas indicated by circles 116, 117, 118, and 119 in FIG. 11 are continuous lines, just as they were in the original image shown in FIG. 10. This process makes it possible to accurately detect the location of the crack based on the shape and position information of the crack recognized by the crack feature recognition unit 22. Furthermore, the thickening process eliminates the fine jagged pattern of the crack, leaving larger shape features. This makes it possible to accurately detect the location of the crack even if there is an error (inaccuracy) in the accuracy of the original hand-drawn crack. It is preferable that the degree to which the width of the lines and dotted lines is to be thickened can be set by visually checking the lines and dotted lines of the binary image obtained by edge detection processing of the image data of the original line drawing, for example, on the display monitor 9.

[0030] FIG. 13 shows a flowchart (subroutine flowchart) of the crack detection process (step S5). The new image data shown in FIG. 8 is preprocessed to convert lines and dotted lines into a series of thick binary data (see FIG. 12 for the area within frame 115). In this binary data, locations containing lines and dotted lines of old cracks in the old image data shown in FIG. 5 (see FIG. 5 for the area within frame 105) are detected by pattern matching (step S51). For cracks whose locations are identified by this pattern matching process (Yes in step S52), the detected locations are confirmed (step S53). On the other hand, if there are cracks whose corresponding locations are unknown (No in step S52), the plan view of FIG. 8 is preprocessed to thicken the plan view, for example, so that the plan view shown in FIG. 12 is displayed with respect to frame 115. Furthermore, the old cracks shown in FIG. 5 are displayed superimposed on this plan view in a distinctive color or line style. The system user visually checks the display monitor 9, operates the pointing device 10 to select a series of linear cracks converted into vector data, and moves and places them in the correct position (step S54). Even when this visual check is performed, the effect of the thickening process can be obtained. If the locations of all the cracks have not been determined (No in step S55), the process returns to step S52. If the locations of all the cracks have been determined (Yes in step S55), the process proceeds to the next step S56.

[0031] In step S56, the series of lines and dotted lines shown in the new image data in Figure 8 are displayed on the display monitor 9 with different line types, line widths, colors, etc. to make them distinguishable, such as thick line parts containing the old crack being black and thick line parts not containing the old crack being white (see Figure 15).

[0032] In step S57, of the series of thick lines showing old cracks in black in the new image data shown in Fig. 8, the old cracks that were in the old image data are shown with thin solid lines, and the other parts, i.e., the newly formed cracks, are shown with thick lines. In this way, the line type, line width, color, etc. are changed to make them distinguishable and displayed on the display monitor 9 (see Fig. 15). After this processing, the process returns to the main routine (step S58).

[0033] FIG. 14 shows a flowchart (subroutine flowchart) of the new crack confirmation process (step S6). In the new crack confirmation process, lines in the new image data that were determined not to be part of a series of thick lines including old cracks in the crack detection process (step S5, see FIG. 13) are identified and set as new cracks. This is because image data obtained by photography may contain not only cracks but also color-coded patterns resulting from slight differences in concrete mix during tunnel construction, which may appear as solid and dotted lines. In step S56 of the crack detection process, the new image data in FIG. 8 is visually inspected to determine whether the area is a new crack (step S61) for areas represented by hollow lines as a series of thick lines after preprocessing that do not include old cracks. If the area is a crack (Yes in step S62), the area is changed to a new crack and displayed with the same thick line as the new crack set in step S57 (step S63). On the other hand, if the area is not a new crack (No in step S62), the line and dotted line are deleted (step S64). This deletion process refers to the case where the corresponding lines and dotted lines in Fig. 8 are deleted from the image data, or, for example, the case where the lines and dotted lines in Fig. 8 are displayed with a thick outline line superimposed on them, or the case where the lines are displayed with a line type, line width, or color that is completely different from the other lines. If the crack detection means (step S5) has not yet checked all of the locations displayed with outline lines (No in step S65), the process returns to step S62. If the check of all locations has been completed (Yes in step S65), the process returns to the main routine (step S66).

[0034] In Figure 15, the information about the old crack that appeared in the old image data is shown in thin lines in the new image data shown in Figure 8, and among the current lines and dotted lines in the part including the old crack, newly formed cracks are shown in thick lines, and further, the lines and dotted lines other than those including the data of the old crack and that are to be subjected to the new crack confirmation process (step S6, see Figure 14) are shown in outline lines 120. If the new crack confirmation process (step S6) determines that the white line 120 is a newly formed crack, it is shown in thick solid lines, and if it is determined that it is not a crack, it is subject to deletion processing.

[0035] As explained above, according to this image processing system, in preprocessing before the crack detection process, edge information detected using new image data is thickened, thereby reducing false detections of no corresponding cracks due to errors caused by hand-drawn accuracy or photographing conditions, and realizing rapid crack detection. As a result, in the new image data captured by the new photograph, old cracks drawn in the old image data are shown with thin lines, and newly formed cracks including the old cracks are shown with thick lines, so the progress of the crack can be seen at a glance.

[0036] The present invention is not limited to the above-described various embodiments, and various modifications are possible without departing from the spirit of the invention. For example, in this embodiment, the lines and dotted lines of cracks in the old image data are converted into vector data. The vectorized cracks can be freely scaled and rotated. For example, even if the scale of the plan view and the developed view of the deformation scanned as the old image data are slightly different or tilted, the line and dotted lines in the new image data are thickened to form a series of thick lines, which has the advantage of facilitating pattern matching. However, if the required accuracy is met, bitmap-format crack data may also be used for pattern matching. In this case, the cracks are grouped using a predetermined pixel matrix and subjected to well-known pattern matching processing. Furthermore, for ease of understanding, the lines and dotted lines in this specification and drawings are shown with different line types and widths for the old crack portions and the new crack portions. However, when displaying in color, different colors may also be used. [Industrial Applicability]

[0037] This idea is effective for recognizing cracks on the inner walls of tunnels, but can also be used to recognize cracks that appear on the walls of buildings and other structures. [Explanation of symbols]

[0038] 1. Image processing system 2 CPU 3 ROM 4 RAM 5 HDD 6. Scanner 7. Camera 8. Printer 9 Display Monitor 10 Pointing Device 21 Input section 22 Crack feature recognition unit 23 Pretreatment section 24 Crack detection unit 25 Output section

Claims

1. an input unit for inputting old image data depicting old cracks on the inspection wall surface and new image data depicting cracks on the inspection wall surface at a time newer than the old image data; a crack feature recognition unit that recognizes the shape features and position of the old crack depicted by the binary image data obtained by performing edge detection processing based on the old image data; a pre-processing unit that recognizes lines and dotted lines from binary image data obtained by performing edge detection processing on the new image data, and thickens the recognized lines and dotted lines by bordering the edges of the binarized image, thereby making adjacent lines and dotted lines into a series of thick lines; a crack detection unit that detects locations containing old cracks from a series of thick lines obtained by processing by the preprocessing unit based on information on the shape characteristics and position of the old cracks (not subjected to thickening processing) recognized by the crack feature recognition unit, and displays the series of thick lines containing the old cracks in the new image data so that they can be distinguished from other lines; an output unit that outputs, displays, or prints out the image data processed by the crack detection unit.

2. 2. The image processing system according to claim 1, Furthermore, the image processing system is characterized in that the crack detection unit indicates in the new image data a series of thick lines including the old crack so that the old crack location can be distinguished from other locations.

3. 2. The image processing system according to claim 1, the output unit includes a display monitor and a pointing device; The image processing system is characterized in that, when multiple locations are detected in the series of thick lines obtained by processing by the pre-processing unit that are judged to include the old crack due to errors in the information on the shape characteristics and position of the old crack, the crack detection unit displays the multiple detected locations on the display monitor, and indicates the correct location specified by the pointing device as the location containing the old crack, and shows the series of thick lines including the old crack in the new image data so that they can be distinguished from other lines.

4. 4. The image processing system according to claim 1, An image processing system characterized in that the old image data is image data obtained by reading a plan view on which old cracks on the inspected wall surface are hand-drawn.

5. 5. The image processing system according to claim 4, The image processing system further comprises a new crack confirmation unit that identifies newly formed cracks and deletes lines that are not cracks for other lines that are determined by the crack detection unit not to be a series of thick lines including old cracks.

6. an input step of inputting old image data depicting old cracks on the inspection wall surface and new image data depicting cracks on the inspection wall surface at a time newer than the old image data; a crack feature recognition step of recognizing the shape features and position of the old crack depicted by the binary image data obtained by performing edge detection processing on the old image data; a pre-processing step of recognizing lines and dotted lines from binary image data obtained by performing edge detection processing on the new image data, and thickening the recognized lines and dotted lines by bordering the edges of the binarized image to make adjacent lines and dotted lines into a series of thick lines; a crack detection process for detecting a location containing a previous crack from the series of thick lines obtained by the preprocessing process based on the shape characteristics and position information of the previous crack recognized in the crack characteristic recognition process, and displaying the series of thick lines containing the previous crack in the new image data so that they can be distinguished from other lines; an output step of outputting, displaying or printing out the image data processed by the crack detection step.

7. A computer-readable image processing program, which, when executed, causes a computer to: an input unit for inputting old image data depicting old cracks on the inspection wall surface and new image data depicting cracks on the inspection wall surface at a time newer than the old image data; a crack feature recognition unit that recognizes the shape features and position of the old crack depicted by the binary image data obtained by performing edge detection processing based on the old image data; a pre-processing unit that recognizes lines and dotted lines from binary image data obtained by performing edge detection processing on the new image data, and thickens the recognized lines and dotted lines by bordering the edges of the binarized image, thereby making adjacent lines and dotted lines into a series of thick lines; a crack detection unit that detects locations containing old cracks from the series of thick lines obtained by processing by the preprocessing unit based on information on the shape characteristics and position of the old cracks recognized by the crack feature recognition unit, and displays the series of thick lines containing the old cracks in the new image data so that they can be distinguished from other lines; and an output unit that outputs, displays, or prints out the image data processed by the crack detection unit.

Images