JP6937642B2 - Surface evaluation method and surface evaluation device - Google Patents

Description

The present invention relates to a surface evaluation method and a surface evaluation device for evaluating the surface state of a structure.

When measuring cracks in concrete structures, it is often measured directly on the concrete surface using a crack gauge or scale.
Further, when measuring cracks or the like in an image such as a photograph, it is necessary to reflect the scale or gauge in the same image as the measurement target at a position where the scale or gauge can be measured (see, for example, Patent Document 1 and Non-Patent Document 1). ). If it is possible to shoot in close proximity, shoot in front of the measurement target and acquire the image. Then, in the techniques described in Patent Document 1 and Non-Patent Document 1, shading correction and binarization processing of the image are performed to extract the cracked region. Next, in the region determined to be a crack, the area of the crack is calculated by multiplying the apparent area obtained by adding the brightness values of each pixel to the total luminance value obtained by the correction coefficient. As a result, it is possible in principle to measure the crack width of 0.2 mm, which is a measurement standard, with sub-pixel accuracy.

In addition, the distance between several laser beam points, left-right angle, vertical angle, objective distance, scale, etc. are obtained and normalized by these parameters to obtain an entire image of the concrete structure and the status of the structure. A technique for informing numerical values is also being studied (see, for example, Patent Document 2).

Japanese Unexamined Patent Publication No. 2003-214827 Japanese Unexamined Patent Publication No. 2004-69434

Nobunori Takeda et al., "Measurement of cracks in concrete structures by image processing", Obayashi Corporation, Obayashi Technical Research Institute Bulletin, No.79, 2015, [online], [Search on August 28, 2017], Internet, <URL: https://www.obayashi.co.jp/technology/shoho/079/2015_079_36.pdf>

In order to measure the surface condition such as cracks in a concrete structure with sub-pixel accuracy, it is necessary to accurately measure the distance to the evaluation target. For example, the distance between the evaluation object and the camera is often obtained by measuring the laser distance. In this method, it is a condition that the evaluation object such as a wall surface and the camera face each other. However, at the evaluation site, it is often difficult to shoot face-to-face. In particular, when it is difficult to move the camera, the camera is shaken up, down, left, and right to take a picture, so that the geometric relationship between the camera and the evaluation object changes for each image, and it is difficult to perform accurate evaluation.

Further, if a distance camera capable of measuring the shooting depth is used, it is possible to measure the distance of all pixels. However, it is difficult to improve the accuracy even by using LiDAR (Light Detection and Ranging) technology, and it is difficult to measure with the accuracy of 0.2 mm / pixel required for crack width measurement, for example.

-The surface evaluation method for solving the above problems evaluates the surface state of the evaluation target surface using a photographed image of the evaluation target surface. In this case, the surface state is recorded in the evaluation model storage device in relation to the resolution and the optical feature amount of the pixel, a reference image for specifying the imaging surface is projected on the evaluation target surface, and the reference image is included. acquires a photographed image of the evaluation target surface, based on the arrangement of the reference image included in the captured image, calculates the resolution of the pixels included in the captured image, using the evaluation model store, the said resolution The surface state of the surface to be evaluated is evaluated based on the amount of optical features of the pixels included in the captured image.

According to the present invention, the surface state of a structure can be evaluated from a distance using a photographed image.

Explanatory drawing explaining the structure of the surface evaluation system in this embodiment. Explanatory drawing of the crack scale model recorded in the crack scale information storage part in this embodiment. The flow chart of the processing procedure in this embodiment. It is explanatory drawing of the processing procedure in this embodiment, (a) is a positional relationship between a wall surface and a camera, (b) is an explanatory view of a photographed image.

Hereinafter, an embodiment in which the surface evaluation method and the surface evaluation device are embodied will be described with reference to FIGS. 1 to 4. In this embodiment, it is assumed that the surface condition (crack) of the concrete structure is measured as the surface to be evaluated.

As shown in FIG. 1, in the present embodiment, the width (crack width) of the crack 31 generated on the wall surface 30 of the concrete structure is measured by using the camera 10, the laser irradiation device 15, and the evaluation device 20.

The camera 10 is a photographing means for photographing a subject and generating a photographed image (color image).
The laser irradiation device 15 irradiates the imaging range of the camera 10 with three laser beams. The three laser beams are arranged at the vertices of a shape known as a reference image (an equilateral triangle in this embodiment).

The evaluation device 20 is a computer system that calculates the crack width based on the captured image. The evaluation device 20 includes a control unit 21 and a crack scale information storage unit 22.

The control unit 21 includes control means (CPU, RAM, ROM, etc.) and performs processes described later ( each process such as an image acquisition step, a resolution evaluation step, a surface evaluation step, etc.). By executing the evaluation program for that purpose, the control unit 21 functions as an image acquisition unit 211, a resolution evaluation unit 212, and a surface evaluation unit 213.

The image acquisition unit 211 executes a process of acquiring a captured image captured by the camera 10.
The resolution evaluation unit 212 executes a process of calculating the resolution of each pixel included in the captured image. The method of calculating the resolution will be described later.
The surface evaluation unit 213 executes a process of evaluating the cracked state included in the captured image.

The crack scale model 220 is recorded in the crack scale information storage unit 22 as information for calculating the crack width. In this crack scale model 220, a luminance feature amount (optical feature amount) is recorded for each crack width for each resolution.

As shown in FIG. 2, the crack scale model 220 is created by using a plurality of images of the same crack scale taken at different focal lengths. Specifically, the relationship between the luminance feature amount and the crack width is calculated using the captured crack scale according to the resolution determined by the focal length. Then, a calibration straight line with the resolution interpolated is calculated for the crack scales taken with different resolutions to create the crack scale model 220.

( Resolution calculation process)
Next, the resolution calculation process will be described. This process is executed by the resolution evaluation unit 212 of the control unit 21. Using the laser beam (vertices A, B, C) projected on the wall surface 30, the geometric relationship between the camera 10 and the wall surface 30 (concrete plane) is derived, and the resolution is calculated for each pixel. In this case, it is assumed that the following conditions are met.
-The wall surface 30 is a two-dimensional plane.
-Lens aberration has been corrected.
-The camera 10, the irradiation position of the laser beam, and the focal length are known.

When the wall surface 30 and the camera 10 are exactly facing each other, the projected laser beam forms an equilateral triangle arrangement in which the length Lt of one side is known. On the other hand, when the wall surface 30 and the camera 10 do not face each other, the shape is distorted without forming an equilateral triangular arrangement. In this embodiment, this distortion is calculated as the geometric relationship between the camera 10 and the wall surface 30.

First, as shown in FIG. 4, consider the camera coordinate system (xyz) and the image coordinate system (uv).
Since the laser beam is output as parallel rays in the z direction of the camera coordinate system, the x components XA, XB, XC and y components YA, YB, YC of the vertices A, B, and C of the triangle are known values. be. The unknown parameters ZA, ZB, ZC are used as the coordinate values PA (uA, vA), PB (uB, vB), PC (uC, vC) of the uv plane of each vertex A, B, C on the image. To calculate.

ここで、ｆは焦点距離である。式（１）より、三角形の頂点座標の３成分を導出することができる。

Here, f is the focal length. From equation (1), the three components of the coordinates of the vertices of the triangle can be derived.

Next, the three-dimensional coordinates of each pixel Pi in the image are calculated. The three-dimensional coordinates can be determined by calculating the intersection of the wall surface 30, the origin O, and the three-dimensional straight line connecting each Pi in the UV plane. The plane of the wall surface 30 can be calculated from the coordinate values of the vertices A, B, and C. First, the vectors AB and AC are as follows.

As a result, the normal vector of the plane can be obtained from the outer product of the vectors AB and AC. The plane equation passing through the vertex A is as follows, where n (l, m, n) is the normal vector obtained by the equation (2).

Next, consider a three-dimensional straight line. For example, the equation for the straight line O-PB passes through the origin O. In order to make it a unit vector, the length L of the unit vector of a straight line is calculated as follows.

Using the length L, the linear equation is as follows.

Putting t in equation (5) to find the intersection gives:

When decomposed by x, y, z, it becomes as follows.

Substituting equation (7) into the planar equation of equation (3) gives the following.

Here, the solution for t is as follows.

By substituting Eq. (9) into Eq. (7), the intersections x, y, and z can be obtained.

From equation (10), the intersection for uB was obtained. Similarly, for each Pi, the intersection, that is, the three-dimensional coordinates can be calculated by the same procedure.
Since the three-dimensional coordinates of each pixel can be obtained, the resolution (mm / pixel) of each pixel can be calculated. Each Pi corresponds to one pixel. Therefore, the resolution Ri of Pi is calculated using the three-dimensional coordinates of two points (Pi-1, Pi + 1) adjacent to each other on both sides of Pi. However, since the resolution may differ between the vertical and horizontal directions, the resolution is the average value of the vertical and horizontal values.

From the above, the distance between the camera and the wall surface can be determined, and the resolutions of all the pixels of one image can be calculated.

(Evaluation processing of cracks)
Next, the processing procedure of the crack evaluation process will be described with reference to FIG.
First, the control unit 21 of the evaluation device 20 executes an image acquisition process (step S1-1).

As shown in FIG. 4A, it is assumed that the crack 31 of the wall surface 30 of the concrete structure is evaluated. In this case, first, the laser irradiation device 15 is used to irradiate the wall surface 30 with three laser beams in a predetermined arrangement (in this embodiment, an equilateral triangle). In this case, bright spots (laser spots) of laser light are generated on the wall surface 30 in the shape of triangles of vertices A, B, and C.

Then, the wall surface 30 of the concrete structure is photographed using the camera 10.
In this case, as shown in FIG. 4B, color photography is performed so that the image includes the crack 31 and the vertices A, B, and C. Then, the captured image 500 is input to the evaluation device 20. In this case, the image acquisition unit 211 of the control unit 21 acquires the captured image 500.

Next, the control unit 21 of the evaluation device 20 executes the process of identifying the cracked region (step S1-2). Specifically, the surface evaluation unit 213 identifies all pixels including cracks in the captured image 500. Here, for example, crack detection is performed by binarization processing using the percolation method described in Non-Patent Document 1, and crack position determination processing is performed by thinning processing.

Next, the pixels specified in step S1-2 are sequentially set as evaluation target pixels, and the following processing is repeated for each evaluation target pixel.
Here, the control unit 21 of the evaluation device 20 executes the resolution calculation process (step S1-3). Specifically, the resolution evaluation unit 212 of the control unit 21 calculates the resolution of the pixel to be evaluated using the above-mentioned equation (11).

Next, the control unit 21 of the evaluation device 20 executes the crack width calculation process (step S1-4). Specifically, the surface evaluation unit 213 of the control unit 21 calculates the brightness of the evaluation target pixel. Next, the surface evaluation unit 213 calculates the crack width from the calibration straight line in the crack scale information storage unit 22 based on the resolution of the pixel to be evaluated and the brightness of the pixel.
The above process is repeated for all pixels containing cracks.

Then, the control unit 21 of the evaluation device 20 executes the output processing of the evaluation result (step S1-5). Specifically, the surface evaluation unit 213 of the control unit 21 outputs the crack width calculated for each pixel in the captured image 500.

According to this embodiment, the following effects can be obtained.
(1) In the present embodiment, the crack scale model 220 for calculating the crack width is recorded in the crack scale information storage unit 22. This crack scale model 2
20 is recorded for each resolution. When the crack is evaluated by the sub-pixel, the crack region and the peripheral region are included in one pixel. In this case, depending on the resolution , even if the cracks have the same width, the ratio of the cracked region to the peripheral region is different, so that the brightness of the pixel including the cracks changes. Here, by using a crack scale model according to the resolution , an accurate surface state can be evaluated.

(2) In the present embodiment, the control unit 21 of the evaluation device 20 identifies the crack region (step S1-2), calculates the resolution (step S1-3), and calculates the crack width (step S1-4). ) Is executed. For each pixel included in the captured image, in accordance with the distance between the wall 30 and the camera 10, the resolution is different. In consideration of this difference in resolution , the state can be accurately evaluated for each pixel containing cracks.

(3) In the present embodiment, color photography is performed so that the image includes the crack 31 and the vertices A, B, and C. By reflecting the vertices A, B, and C generated by laser light irradiation in the captured image, the positional relationship between the wall surface 30 and the camera 10 can be grasped, and the size of cracks, etc. in one captured image can be grasped. The surface condition of the surface can be evaluated efficiently.

Moreover, the said embodiment may be changed as follows.
-In the above embodiment, the crack width is calculated using the evaluation device 20. It is not necessary to perform all by computer processing, and a part of the calculation may be performed by manual calculation or the like.

-In the above embodiment, it was applied to the calculation of the crack width of the wall surface of the structure. The application target is not limited to the evaluation of cracks on the wall surface as long as the amount of optical features such as brightness changes depending on the surface condition. For example, it can also be used for evaluation of scratches and the like on the surface to be evaluated.

-In the above embodiment, a crack scale model in which the luminance feature amount, the crack width, and the resolution are variables is used. The variables used in the crack scale model are not limited to these. For example, the brightness of the ambient light may be further used as a variable. In this case, the illuminance of the ambient light is changed, a crack scale model is generated using the captured image of the crack scale, and the crack scale information is recorded in the crack scale information storage unit. Then, at the time of evaluation of cracks, the brightness of the ambient light is acquired by an illuminance meter or the like. Then, using the crack scale model corresponding to this illuminance, the crack width is calculated from the resolution and the luminance feature amount.

-In the above embodiment, the positional relationship between the wall surface 30 and the camera 10 is calculated using the laser light arranged in an equilateral triangle as a polygon. The arrangement of bright spots of a plurality of laser beams that can specify a two-dimensional plane is not limited to an equilateral triangle. It is also possible to use polygons of other shapes, cross arrangements, and the like.

Further, when the wall surface 30 is a curved surface, a divided region (for example, a mesh-like projection) formed by a plurality of polygons may be used as in a grid pattern. In this case, the positional relationship between the wall surface 30 and the camera 10 is calculated using a polygon (divided region) including cracks.

10 ... camera, 15 ... laser irradiation device, 20 ... evaluation device, 21 ... control unit, 22 ... crack scale information storage unit, 211 ... image acquisition unit, 212 ... resolution evaluation unit, 213 ... surface evaluation unit, 30 ... wall surface, 500 ... Photographed image.

Claims (4)

It is a method of evaluating the surface state of the evaluation target surface by using a photographed image of the evaluation target surface.
Using a plurality of images of the same crack scale taken at different focal lengths, a crack scale model was created in which the surface state was associated with the resolution and the optical features of the pixels according to the focal length.
A reference image for identifying the photographing surface is projected onto the evaluation target surface, and the image is projected.
The photographed image of the evaluation target surface including the reference image is acquired , the cracked region is specified, and the cracked region is identified .
Based on the arrangement of the reference image included in the captured image, the resolution of the pixels included in the captured image is calculated.
On the basis of the position of the reference image included in the captured image, for each pixel of the cracked area, it calculates the resolution of each pixel of the crack region included in the captured image,
For each pixel of the cracked region using the crack scale model, and resolution of each pixel, on the basis of the optical characteristic of each pixel, a surface assessment and repeating the evaluation of the surface condition of the cracked region Method. The surface evaluation method according to claim 1, wherein the reference image is composed of polygonal vertices generated by laser light irradiation. Based on the light feature amount before Symbol cracking, surface evaluation method according to claim 1 or 2, characterized in that calculating the crack width as the surface state. An evaluation model storage device that records a crack scale model in which the surface state is associated with the resolution according to the focal length and the optical feature amount of the pixel using a plurality of images of the same crack scale taken at different focal lengths.
An evaluation device including a control unit that evaluates the surface state of the evaluation target surface using a photographed image of the evaluation target surface.
The control unit
A photographed image obtained by projecting a reference image for specifying the imaged surface on the evaluation target surface is acquired, and a cracked region is specified .
Based on the arrangement of the reference image included in the captured image, the resolution of the pixels included in the captured image is calculated.
On the basis of the position of the reference image included in the captured image, for each pixel of the cracked area, it calculates the resolution of each pixel of the crack region included in the captured image,
For each pixel of the cracked region using the crack scale model, and resolution of each pixel, on the basis of the optical characteristic of each pixel, a surface assessment and repeating the evaluation of the surface condition of the cracked region Device.

Images