JPH01291382A - Image reading method - Google Patents

Abstract

PURPOSE:To shorten the time required for read of an image, an image processing and an analysis by executing read of image information with regard to a designated read minute visual field along a virtual mesh line. CONSTITUTION:On a picture image which has photographed a crack A, for instance, by a CPU (central arithmetic unit), a mesh line l is formed virtually at a desired interval and in a relative position, and a picture is read successively at every minute visual field B along the virtual mesh line l. Accordingly, the image analysis is executed only with regard to a designated minute visual field, that is, a limited read area. In such a way, the analysis time can be shortened. Also, although the accuracy becomes lower than data related to the crack A at the time when all areas of the crack picture image are read the analysis has been executed, this method is effective in case when it is desired to obtain quickly rough data related to the crack A.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image reading method, in particular, by photographing cracks occurring on the surface of an object such as concrete I or a structure, and reading this photographic image. This invention relates to a method of photoelectrically reading a photographic image when cracks are analyzed by photoelectrically reading the photographic image and performing image processing and analysis.

[Conventional technology] Traditionally, cracks that occur on the surfaces of civil engineering and architectural structures such as buildings, bridges, roads, and embankments cause water to leak, corrode the steel frames, and accelerate the deterioration of these structures. .

In particular, cracks that have recently appeared in exterior materials such as concrete have caused them to fall off, causing serious accidents and becoming a social problem.

Up until now, the inspection and VN analysis method for cracks on the surface of these structures has relied heavily on visual inspection.
Apply a crack scale to the crack location at the site, read the crack width, and grasp the crack shape by sketching. Then, by aggregating and analyzing this data for each virtually set unit area, the future progress of cracks, etc., was predicted, and the timing of repairs, etc. was determined.

Alternatively, when reading the width of a crack, etc., draw mesh lines (orthogonal grid mesh) on the actual measurement surface at predetermined equal intervals in both the X and Y directions, and visually check the mesh lines. The width of the crack was measured at the point where the crack crossed.

[Problem to be solved by the invention] However, this visual measurement of cracks not only requires a great deal of time and effort and is extremely inefficient, but also presents difficulties in terms of accuracy due to individual differences among the measurers. was there.

In order to eliminate these drawbacks of visual inspection, the present inventors first took a photograph of the surface to be inspected such as a structure at the site, and then obtained a photoelectrically converted image signal from this photographic image. We considered a crack measurement device that performs crack analysis by signal processing using an image processing and analysis device. In this device, the entire field of view of a photograph is sequentially read and scanned into a minute area fi of a predetermined area limited by the size of the light-receiving surface of the photoelectric conversion element or the processing capacity of the image processing and analysis device. Performing repeated analysis. .

The micro area can be set to a square of, for example, 0.5 mm x 0.5 mm. is divided into 104 minute areas and read.

However, in a general image processing analysis device, it takes a relatively long time to analyze the minute area, so a large amount of intercostal space is consumed in the entire field of the film.

The present invention has been made in view of the above-mentioned circumstances, and it is an object of the present invention to reduce the time required for reading and analyzing photographs in a device that performs image processing and analysis from photographic images. 1. An object of the present invention is to provide an image reading method that can achieve i7 reduction.

[Means for Solving the Problems] The above-mentioned object of the present invention is to divide a photographic image of a crack in an object into regions of a number of phases, photoelectrically read the image of the region fσ and convert it into information. In an image reading method of an analysis process of a crack photographic image, which analyzes information to obtain data regarding cracks, and synthesizes the obtained data to create data regarding cracks of the entire photographic image or a predetermined part, the crack photograph This is achieved by virtually creating a mesh line in the image and reading the crack photographic image region by region and sequentially along the mesh line, in particular by optically magnifying projection or electronic display. This is achieved by displaying the image superimposed on the mesh line and photoelectrically reading the photographic image of the crack based on this display.

[Function] According to the above-mentioned means, image information is not read for the entire area of a photographed photograph, but is performed for a reading minute field of view specified along the virtual mesh line, so that the reading area is It is possible to significantly shorten the image reading time and the image processing/analysis time.

Therefore, although the accuracy is lower than data on cracks in the cracks that is obtained by reading and analyzing the entire area of a photo image of a crack, it is an effective image reading method when you want to quickly obtain general data about cracks in a photo image. .

[Example] Hereinafter, the present invention will be explained in detail with reference to the drawings.

FIG. 1 is a block diagram illustrating a crack measuring device constructed based on the image reading method of the present invention.

The crack measuring device is roughly divided into an 11-piece unit 1 and a photographic image reading/analysis/output unit 10.

The photographing unit 1 can be used, for example, to produce high-resolution black-and-white copying [1 silver halide film, such as a silver halide photosensitive material for high-resolution color photography, and an optical camera that uses it. Photograph the surface of the structure, building, etc. to be inspected for cracks.

The photographic film used to project the cracks may be of any size, but a general and easy-to-use photographic film, such as blown film (screen size 50mm x
50 mm), and using this, 1 m x 17
It is possible to photograph 11 inspection target surfaces.

The above-mentioned reading/analysis/output 2, knit 10+ soil, further photographic image reading device 11, image processing]! 11・Analysis device 12, C
P LJ (central processing unit) 13 and display/output splitting 11
j++, and each device is controlled by c p U 13.

FIG. 2 illustrates the photographic image reading device 11, which is a main element of the present invention.

This photographic image reading device 11 includes a stage 100 on which a photographed photograph is placed, an X-axis axis 103 and a Y-axis axis 104 that move the stage 1oo in the X-axis and Y-axis directions.
, a microscope 105 for observing the photograph on the stage, a microscope 105 for viewing the photograph as an LRD image (including an area image scanner for photoelectric conversion), and a microscope 105 for viewing the photograph at the position of the photographic enlarging device 101. The X-axis cylinder 2 and -9108 are moved between the positions of 10 and 10!i.

The photo image reading device 11 is also equipped with a photo enlarger 101 and an XY digitizer 102, and can also enlarge and project the photo on the stage 100 onto the XY digitizer 102 by the enlarger 101. J: The sea urchin is turning.

FIG. 3 shows an example of a projected image of an enlarged and projected crack photograph, and shows that crack images Δ are irregularly distributed on the projected image 210. Here, for convenience of explanation, the projected image 210 is shown with a virtual mesh line p superimposed on the projected image. The virtual mesh line I is not actually drawn on the photograph 210, but a line or an image displaying the virtual mesh line is superimposed on the photograph 210, so that it can be visually seen as if a line had been drawn. It's fine.

As a method for adding virtual mesh line force to a photographic image, for example, an XY digitizer 10 attached to a photographic enlarger 101 can be used.
There is a method of superimposing mesh lines made of orthogonal lattices on a transparent PLUSDEC sheet (not shown) on an optically enlarged projected image of a photograph on two sheets of paper. Other methods include optically superimposing a virtual mesh line image consisting of orthogonal grids displayed on a CRT (not shown) on an optically enlarged projection image of a photograph, or overlapping the entire field of view of a photograph. Image sensor (
There is a method of electronically superimposing and displaying a virtual mesh line image on a 1Q monitor image that has been read in (not shown). The virtual mesh line β is used to perform virtual partitioning (the shape is generally a square, but it can also be set in the shape of a rectangle) when organizing the read data of the crack △. , the interval is the actual size of the surface to be measured, for example, 1
0.30.50.100 mm and other spacings.

For example, the operator superimposes the transparent plus deck seam 1 with mesh lines on the crack photographic image, and while looking at the superimposed pattern, determines the interval between the virtual mesh lines ρ, taking into consideration the shape and quantity of the cracks Δ. The operator can set the relative positions of the and photographic images, or determine the spacing and relative positions of the virtual mesh lines ρ as appropriate without creating a rock pattern, and select or select these on a pre-programmed menu screen. They can be set by inputting (direct input) using the keyboard, or they can be set using predetermined values without the operator determining the spacing of the virtual mesh line ρ and its relative position to the photographic image. You may also convert them to Progera l\.

When reading the crack Δ, out of the read minute field of view (read minute area), read only the minute field of view B along the position information of the virtual mesh line, Q, as shown in the circular enlarged part in Figure 3. The photographic image is read according to the programmed program.

The centers of the reading minute fields of view B are on the virtual mesh line in both the X and Y directions, and the respective reading minute fields of view B are set to be adjacent to each other as shown in the figure. This setting work is performed by the CPU 13 according to a program installed in advance.

XY digitizer 102 (well, specify the relative position of the above virtual mesh line 9 drawn on the transparent plus book sheet when superimposed on the optically enlarged projection photo image with the cursor -c, read and register in the CPU) The operator moves the virtual mesh line superimposed on the projected photo image to Sea 1, performs positioning, and then presses the cursor button between 10 and 2. In other words, according to this embodiment, the reading area is determined according to the interval between the virtual mesh lines β, and the reading time is significantly reduced compared to the case where the entire photograph 210 is read. The reduction can be covered.

The CPU 13 controls the X-axis motor 103 based on the distance between the virtual mesh lines and the relative position information with respect to the photographic image.
, and controls the Y-axis motor 104.

, and move the micro area on the photograph corresponding to the position coordinates of the mesh line to the observation position of the microscope 105 and the TV camera 106.
placed at the imaging position.

Furthermore, when the relative position of the virtual mesh line ρ is input using the XY digital tile "102", the CPU 13
By controlling the rotation amount of the axle box 2 motor 108,
03, it works together with the Y-axis motor 104 to automatically correct the predetermined position of the photographic image so that it is on the optical axis of the microscope 105.

The minute area placed under the objective lens 107 of the microscope 105 is observed by the microscope, photoelectrically converted by the TV camera 106, and displayed on an image display (not shown).

Note that the minute area is limited by the size of the area image sensor or the processing capacity of the image processing and analysis device, so for example, the area equivalent to 250,000 pixels is 0.5g
It is set to a square of 0.51gm.

In addition, to enlarge a photographic image onto an XY digital camera, it is possible to divide all the two sides of one frame and project it in -C enlargement.There are several numbers of divisions, which can be selected as appropriate.

Therefore, the sub-fret corresponding to the film position of each divided area of the divided enlarged projection ++, ', is automatically corrected by the CPU 13 based on the coordinate tl calculation, and control the amount of rotation of the X-axis box 2 motor 108 for
The X-axis motor 103 and the Y-axis motor 104 work together,
A predetermined portion of the photographic film is automatically placed on the optical axis of the microscope 105 by 16 degrees.Also, during actual reading, when a photographic film holder (not shown) is conveyed to the microscope objective lens, Another filler holder can be placed in place on the enlarger. Then, a copy of the photograph to be analyzed is attached to this other film, and if necessary, the area corresponding to the area to be enlarged and read by the microscope and its surrounding areas are enlarged and projected using an enlarger for visual inspection. It can be configured to allow observational monitoring. In such a configuration, registration between the original image and the duplicate image is performed by an operator. In this registration, only the rotation is important, and positional deviations in the X direction and the X direction can be corrected by moving the observation area of the enlarged projection image of the enlarger in the X direction and the X direction, respectively.

The present invention forms a virtual mesh line on a photographic image of a crack, and reads along the virtual mesh line in a micro field of view B.
In this case, the total analysis time can be significantly shortened compared to the case where the entire area of the photograph 210 is read and analyzed by setting and limiting the reading part of the crack A.

Therefore, returning to FIG. 1, the reading device 11 reads minute regions along the virtual mesh line for each minute = 13- region, and each time one is read, it is analyzed by the analysis device to determine the width of the crack. , angle, etc. are obtained, this is repeated n times, and the results are collected by the CPU 13 and the entire crack analysis results are output from the display/output device 14.

However, n is the total number of micro regions (total number of micro fields of view) in which the photograph is read along the virtual mesh line for each micro region. n
The finer the virtual mesh line, the more, and conversely, the coarser the line, the fewer.

Note that the output display is a crack width distribution histogram, a crack diagram, a crack angle distribution histogram, etc.

FIG. 4 shows the sequence 1 to 1 calls sent to the CPU of the crack measuring device.
It is.

First, at block 200, the operator specifies a reading mode. In this reading mode, in the present invention, a virtual mesh line β is formed and reading areas are sequentially specified along the virtual mesh line Q, but the interval between the virtual mesh lines and the relative position with respect to the photographic image is displayed. Also included are numerical die reflexes 1-in specifying selections, desired spacing and relative positions on both sides of the menu that are pre-programmed to do so.

Another reading mode is to not set virtual mesh lines,
The entire photograph is unconditionally read minute by minute area.

Therefore, the CPU 13 determines the input reading mode, and if it is the virtual mesh line mode, determines the interval and relative position of the virtual mesh line ρ based on the mode, and determines the reading micro field of view along the virtual mesh line. A command is issued to the reading device to read B as shown in block 201.

It should be noted that this crack measuring device is also capable of reading the entire area of the photograph in minute areas, and in that case, block 2021 instructs the reading device to read the entire area unconditionally. Next, in block 203, an instruction is given to capture image information for each minute area within the specified area. Block 2
At step 04, the image processing and analysis device is instructed to start processing and analyzing the captured image information.

The analysis results such as the width, length, position, direction, etc. of the crack are sucked up from the image processing and analysis device in block 205, and then in block L1 20G, the analysis results of the entire reading area are sent to the CPU in the J mm specified area. The presence or absence of a reading area is determined. If there is an unread area, block 203
The control returns to /) (repeated). When the analysis of all the reading areas is completed, they are synthesized and organized in block 207, and the analysis results are printed out or displayed on a display in blocks 208 and 20'3.

[Effects of the Invention] As described above, according to the image reading method of the present invention, a CPU (central processing unit), for example, can virtually scan a photographic image of a crack at desired intervals and relative positions. A mesh line is formed in the area, and photographs are read sequentially for each minute field of view along this virtual mesh line, so image analysis is performed only for the specified minute field of view, that is, a limited reading area, and the photograph is read. [Compared to image analysis, the analysis time can be significantly reduced1 for the entire area. Therefore, the accuracy is lower than the data related to cracks obtained by reading and analyzing the entire area of a photographic image of cracks, but if you want to quickly obtain general data regarding cracks in a photographic image, the present invention method is effective.

Also, crack the virtual mesh line photo i1! By projecting or displaying on a monitor superimposed on the i-image, the operator can know what part of the photographic image is to be read, and it is easy to compare the measurement results with the conventional method.

[Brief explanation of the drawing]

FIG. 1 is a diagram illustrating a crack measuring device configured based on the image reading method of the present invention, FIG. 2 is a block diagram of a photographic image reading device in the device, and FIG. 3 is a photographic image of a crack and a virtual mesh. Line superimposition pattern diagram, 4th
The figure shows the CP in the 5-attenuation image reading/analysis/output unit 1-
FIG. 2 is a flow chart showing U's sequence control 1 to roll. FIG. [Numbers in the diagram] 1: Photographing unit 1- 10: Reading/analysis/output unit 1-11: Photographic image reading device 12: Image processing/analysis device 13: CPU 14: Display/output device 10
0・Stage 1o1. Photographic enlarger 102
・XY digitizer 1031XNl + motor 104
:Y-axis motor 105・Microscope 106:TV camera 107. Objective lens 108: X-axis box 2 motor 210° Superimposed pattern of crack photographic image and virtual mesh line Diagram A: Crack B: Reading minute field of view fJ: Virtual mesh line area Attorney Patent Attorney (8107) Kiyotaka Sasaki (
3 idiots) = 18-

Claims (2)

[Claims] (1) Divide a photographic image of a crack in an object into multiple regions, photoelectrically read the image of each region and convert it into information, analyze this information to obtain data regarding the crack, and use the obtained data to In an image reading method for an analysis process of a cracked photographic image, which synthesizes data on the entire photographic image or on cracks in a predetermined part, a mesh line is virtually provided within the cracked photographic image, and the cracked photographic image is An image reading method characterized by sequentially reading each area along the mesh line. (2) displaying the crack photographic image superimposed on the mesh line by optically enlarged projection or electronic display;
2. The image reading method according to claim 1, wherein the crack photographic image is read photoelectrically based on this display.