JP7649194B2 - Striatal Imaging Diameter Measurement System - Google Patents

Description

This invention relates to a streak image diameter measurement system that uses image data to measure the diameter of the streak.

Patent Document 1 discloses a wire rope inspection device that uses a camera to capture an image of a wire rope and calculates the actual diameter of the wire rope using the image obtained by capturing the image. A conversion coefficient (actual dimension per pixel) between the number of pixels in the width of the image area of the wire rope captured by the camera and the actual width of the wire rope is set, and the actual width of the wire rope is calculated from the image of the wire rope using this conversion coefficient.

Patent No. 5769875

When calculating the actual width of a wire rope from the number of pixels in the width of the image area of the wire rope, the relationship between the camera's posture at the time of imaging and the longitudinal direction (traveling direction) of the wire rope must be correct. In other words, if the horizontal (row) arrangement of the imaging elements of the camera used to image the wire rope is correctly perpendicular to the longitudinal direction (traveling direction) of the wire rope, the actual width of the wire rope can be correctly calculated from the number of pixels in the width of the image area of the wire rope. However, if the horizontal arrangement of the camera's imaging elements is not perpendicular to the longitudinal direction of the wire rope, a value greater than the wire rope's width will be calculated. The wire rope inspection device described in Patent Document 1 requires high precision in the installation posture of the camera that images the wire rope.

The purpose of this invention is to make it possible to measure the diameter of a filament with high accuracy without having to pay much attention to the position of the camera that images the filament.

The invention also aims to improve the accuracy of measuring the diameter of the umbilical cord.

The present invention further aims to provide a measurement system suitable for measuring the diameter of a filament having a twisted structure.

The image diameter measurement system for a filament according to the present invention comprises a camera including an image sensor that images the filament so that the entire width direction is captured and outputs image data including the filament in the subject image, a lighting device disposed on the opposite side of the camera across the filament, and an image processing device that processes image data provided by the camera, and the image processing device comprises an edge position determination means that determines the edge positions of both ends of the silhouette image of the filament for each line image constituting a frame image represented by the image data provided by the camera, a center line determination means that determines a midpoint position equidistant from the edge positions determined for each line image by the edge determination position means and determines the center line of the filament silhouette image included in the frame image using the determined midpoint position, and a diameter determination means that determines the length between the edges of the filament silhouette image along a line segment perpendicular to the center line for each line image using the edge positions, midpoint position, and the inclination of the center line.

The camera and the lighting device are positioned facing each other with the wire, which is the object to be measured, in between. This positioning allows a silhouette image of the wire to be formed on the image sensor of the camera. The camera is positioned so that the entire width of the wire is captured.

The image sensor may be an area sensor in which many image sensors are arranged two-dimensionally in the row and column directions, or a line sensor in which many image sensors are arranged two-dimensionally in the row direction. In either case, a rectangular frame image is constructed from the image data output from the image sensor in the camera. The frame image includes a silhouette image of the streak that includes the entire streak in the width direction.

For each line image (a horizontal image represented by image data output from a line of multiple image sensors) that constitutes the frame image, the positions of both ends of the silhouette image of the filament are determined, and a midpoint position equidistant from both end edge positions is determined. The center line of the silhouette image of the filament included in the frame image is determined using the midpoint position determined for each of the multiple line images.

The center line of the striatal silhouette image can be determined by connecting the midpoint positions determined for each of the multiple line images. To determine the center line more precisely, the center line may be a regression line calculated by the least squares method using the midpoint positions determined for each line image.

If the longitudinal direction of the wire is perpendicular to the horizontal direction of the image sensor in the camera (the direction in which multiple image sensors are arranged), the center line will match the vertical direction of the frame image. If the longitudinal direction of the wire is not perpendicular to the horizontal direction of the image sensor in the camera, the center line will not match the vertical direction of the frame image and will be tilted.

According to this invention, the length between both edge ends of the filament silhouette image is determined for each line image along a line segment perpendicular to the center line using the end edge positions, midpoint position, and the inclination of the center line. In other words, by using the inclination of the center line of the filament silhouette image, the width (diameter) of the filament silhouette image is determined along the line segment perpendicular to the center line, i.e., along the width direction of the filament silhouette image (the direction perpendicular to the longitudinal direction of the filament silhouette image), rather than simply the length connecting the end edge positions.

Specifically, the cosine or sine of the length connecting the two edge positions can be found using the inclination of the center line (whether cosine or sine is used is determined by whether the reference for the inclination of the center line (zero inclination) is vertical or horizontal). Even if the striatal silhouette image is captured at an angle, the length between the two edge positions can be determined correctly along the width direction of the striatal silhouette image.

The calculated length between both ends is based on the silhouette image of the filament represented by the image data. The diameter (actual size) of the filament corresponding to the length between both ends is calculated according to the pixel resolution (dimension per pixel). The pixel resolution can be uniquely determined from the field of view size and the number of pixels of the image sensor.

In this way, with this invention, the diameter of the filament can be measured correctly even if the image of the filament is taken at an angle. There is no need to pay much attention to the position of the camera that takes the image of the filament, which is particularly convenient when measuring the diameter of a filament that is in use.

Preferably, the diameter determination means includes a radius calculation means for calculating the left and right radii of the filament silhouette image using each of the end edge positions and the midpoint position. The left and right radii can be calculated along a line segment perpendicular to the center line, i.e., correctly along the width direction of the filament silhouette image. By measuring the left and right radii separately, the diameter of the filament can be measured more accurately.

In one embodiment, the diameter determination means determines the length between the two end edges as the sum of the maximum values of the left and right radii calculated for each line image by the radius calculation means. This makes it possible to determine the diameter of the circumscribed circle when the cross section of the filament is not a perfect circle.

In one embodiment, the edge position determination means determines the centroid position of the peak represented by the amount of change in luminance of horizontally adjacent pixels in the line image as the end edge positions. The end edge positions can be determined finer than the pixel unit, and the measurement accuracy of the umbilical cord diameter can be improved even if the image sensor has a small number of pixels (low pixel resolution).

In another embodiment, the edge position determination means determines the positions that show the maximum value of the luminance change amount of horizontally adjacent pixels of the line image as the end edge positions. The end edge positions can be easily determined without obtaining the center of gravity position described above.

When the filament is a wire rope formed by twisting together multiple strands, the diameter calculation means preferably uses an image portion including one peak of the silhouette image of the wire rope included in the frame image to determine the length between both end edges of the filament silhouette image along a line segment perpendicular to the center line. When the silhouette image includes multiple peaks (when the frame image includes a wire rope with a length equivalent to multiple pitches), it is possible to avoid discrepancies between the calculated positions of the left radius and the right radius.

The present invention also provides a device and program for measuring the image diameter of a filament suitable for the above-mentioned filament image diameter measuring system.

FIG. 1 is a block diagram of a wire rope diameter measurement system. A frame image is shown. 1 is a flow chart showing an image processing algorithm for measuring diameter. 11 is a graph showing the luminance and the amount of change in luminance of a line image. The principle of calculating the diameter of a wire rope is shown below. 1 is a graph showing the results of measuring the diameter of a wye rope.

Figure 1 is a block diagram showing a schematic diagram of a wire rope diameter measurement system.

The wire rope 1 diameter measurement system includes a digital camera 11, a lens 12 provided in front of the digital camera 11, an illumination source 13, and a recording/analysis device 14.

An illumination light source 13 is placed on one side of the wire rope 1, which is the object to be measured, and a digital camera 11 and a lens 12 are placed on the other side. The light emitted from the illumination light source 13 is partially blocked by the wire rope 1, and is received by the image sensor (not shown) of the digital camera 11 via the lens 12. The wire rope 1 may be running along its length or may be stationary.

A telecentric lens is preferably used for the lens 12. By using a telecentric lens, only light incident parallel to the optical axis of the lens 12 is focused on the image sensor of the digital camera 11, making it possible to obtain an undistorted image of the subject.

The image sensor of the digital camera 11 may be an area sensor with multiple image pickup elements arranged two-dimensionally in the vertical and horizontal directions, or a line sensor with multiple image pickup elements arranged horizontally. The image pickup elements may be C-MOS (Complementary Metal-Oxide Semiconductor) elements or CCD (Charge-Coupled Device) elements. The frame rate of the image sensor can be set arbitrarily depending on whether the moving wire rope 1 or the stationary wire rope 1 is to be imaged, and further depending on the traveling speed of the moving wire rope 1. For example, the wire rope 1 is imaged at a frame rate of about 10 to 500 fps. The number of pixels of the image sensor is arbitrary, but the more pixels there are, the more detailed (finer) the edge position described below can be determined.

The image data output from the digital camera 11 is sent to the recording/analysis device 14. The recording/analysis device 14 is a computer device equipped with a CPU (Central Processing Unit) that controls the overall operation of the recording/analysis device 14, a memory that temporarily stores various data, a storage medium that stores various data, a communication device for data communication with the digital camera 11, and a display device for displaying the analysis results (graphs showing the measurement results of the diameter d of the wire rope 1) described below, etc., and the image data sent from the digital camera 11 is recorded in the storage medium provided in the recording/analysis device 14, and the diameter d of the wire rope 1 is measured using the image data recorded in the storage medium. The storage medium of the recording/analysis device 14 records the image data sent from the digital camera 11, as well as a program for causing the recording/analysis device 14 to execute the image processing described below, an operating system program, etc.

Figure 2 shows an example of an image (hereinafter referred to as a frame image) 20 represented by one frame of image data captured by the digital camera 11 and recorded on a storage medium provided in the recording/analysis device 14.

As described above, the lighting source 13 is installed on one side of the wire rope 1, and the digital camera 11 is installed on the other side, so that the frame image 20 captured by the digital camera 11 includes a dark image portion 21 (hereinafter referred to as a wire silhouette image) where the light emitted from the lighting source 13 is blocked by the wire rope 1, and a bright image portion around it. As shown in Figure 2, the digital camera 11 is installed so that the entire width of the wire rope 1 is captured.

If the longitudinal direction of the wire rope 1 and the horizontally aligned image sensors of the digital camera 11 are correctly perpendicular when capturing the image, the wire silhouette image 21 will be aligned vertically in the frame image 20. However, if there is a misalignment between the longitudinal direction of the wire rope 1 and the horizontal alignment of the image sensors of the digital camera 11, the wire silhouette image 21 will be tilted in the frame image 20, as shown in Figure 2.

The wire rope diameter measuring device of this embodiment has an algorithm that can correctly measure the diameter of the wire rope 1 even when using a frame image 20 in which the longitudinal direction of the wire rope 1 and the horizontal imaging element arrangement of the digital camera 11 are not correctly perpendicular as shown in Figure 2, resulting in an inclined wire silhouette image 21. The details of the algorithm for measuring the diameter of the wire rope 1 are explained below.

Figure 3 is a flow chart showing an image processing algorithm for measuring the diameter d of the wire rope 1. The process shown in Figure 3 is executed by the recording/analysis device 14. Figure 5 shows the principle of calculating the diameter d of the wire rope 1, and will be referred to as appropriate in the following explanation.

As described above, image data output from the digital camera 11 is recorded on the storage medium of the recording/analysis device 14. The processing described below is performed on the image data recorded on the storage medium.

First, the basic concept of measuring the diameter of the wire rope 1 will be briefly explained. The wire rope 1 is generally made by twisting together multiple strands, so that a spiral valley (groove, nip) is formed at the boundary between adjacent strands, and a peak (crown) is located between the valleys. In measuring the diameter d of the wire rope 1, the diameter of the circumscribed circle of the cross section of the wire rope 1 along the peak of the wire rope 1 is used. As will be explained below, in the process of measuring the diameter d of the wire rope 1 using image data executed by the recording/analysis device 14, the image portion corresponding to the peak of the wire rope 1 is also processed to be used in calculating the diameter d.

The frame image 20 includes a number of line-unit images (hereafter referred to as line images). First, each of the multiple line images included in the frame image 20 is processed, and the left edge position (coordinate _xL ) and right edge position (coordinate _xR ) of the wire silhouette image 21 are determined for each line image (step 31).

Figure 4 is a graph showing the brightness and brightness change of one line of a line image contained in a frame image 20, with the vertical axis of the upper graph representing brightness and the vertical axis of the lower graph representing the brightness change. The horizontal axis in both cases represents the horizontal position of the line image. The actual size of each pixel of the imaging element is known, and the horizontal axis (x coordinate) of the graph shown in Figure 4 is converted to the actual size value.

As described above, the digital camera 11 is positioned so that the wire silhouette image 21 (both the left and right edges) is completely included in the frame image 20. The multiple line images that make up the frame image 20 include low-luminance areas that represent the wire silhouette image 21, and high-luminance areas adjacent to the left and right of the low-luminance areas.

When the difference in brightness between the image data output from horizontally (left and right) adjacent imaging elements is calculated as the brightness change amount, the graph shown in the lower part of Figure 4 can be drawn from the graph shown in the upper part. More specifically, the brightness change amount is calculated by subtracting the brightness value of the left imaging element from the brightness value of the right imaging element, using the brightness values of the two image data output from each of the two horizontally adjacent imaging elements. If the brightness value of the right imaging element is smaller than the brightness value of the left imaging element, the calculated brightness change amount will be a negative value. Conversely, if the brightness value of the right imaging element is greater than the brightness value of the left imaging element, the brightness change amount will be a positive value. In the brightness change amount graph, a peak P1 with a negative value corresponding to the left edge of the wire silhouette image 21 and a peak P2 with a positive value corresponding to the right edge appear.

Using the brightness change amount described above, the left and right edge positions of the wire silhouette image 21 contained in the line image are determined.

The left edge position _xL is calculated by the following equation 1.

Here, f(x) is a function that represents the curve of peak P1 when the horizontal position of the line image is x, and a and b are arbitrary horizontal positions that sandwich peak P1 on both sides. The x coordinate of the center of gravity of peak P1 is calculated by Equation 1, and this is treated as the left edge position _xL of the line image.

The right edge position _xR of the line image can also be calculated using a function that represents the curve of peak P2 in the same manner as in Equation 1 above.

The left and right edge positions _xL , _xR calculated by Equation 1 can be calculated in units smaller than one pixel, for example, in 1/10 to 1/100 pixel units (subpixel units) (the exact pixel resolution is uniquely determined by the distance (field of view size) between the subject wire rope 1 and the digital camera 11 and the number of pixels of the image sensor), and the left and right edge positions (coordinates) _xL , _xR can be determined with high precision.

Instead of taking the centers of gravity of the peaks P1, P2 of the luminance change as the left and right edge positions _xL , _xR , the x coordinates at which the luminance change is maximum may be taken as the left and right edge positions _xL , _xR, respectively. Since there is no need to calculate the centers of gravity, the left and right edge positions _xL , _xR can be easily determined.

Once the horizontal position _xL of the left edge and the horizontal position _xR of the right edge have been determined, the coordinates of the midpoint equidistant from these two positions _xL and _xR are then determined as the midpoint position _xM (step 32).

The above-mentioned processing is performed for each of the multiple line images contained in the frame image.

Once the left edge position _xL , right edge position _xR , and midpoint position _xM have been calculated for all line images included in the frame image 20, a regression line is calculated using the midpoint position _xM calculated for each line image, and this regression line is treated as the center line C of the wire silhouette image 21 included in the frame image 20. The least squares method is typically used to calculate the regression line (step 40).

By calculating the regression line (center line C), the inclination θ of the center line C of the wire silhouette image 21 relative to the frame image 20 is grasped. With reference to FIG. 5, in this embodiment, the inclination θ of the center line C is the angle from the vertical direction, with the vertical direction of the frame image 20 as the reference (inclination zero).

By calculating the left edge position _xL and the right edge position _xR for all line images included in the frame image 20, edge curves similar to sine curves are drawn at the left edge position and the right edge position of the wire silhouette image 21. As described above, the wire rope 1 is made by twisting together a plurality of strands, and peaks and valleys appear repeatedly on the surface of the wire rope 1 along the longitudinal direction of the wire rope 1.

The wire silhouette image 21 shown in Figure 2 has three peaks 21L1, 21L2, and 21L3 on the left edge and three peaks 21R1, 21R2, and 21R3 on the right edge. When the wire silhouette image 21 contains multiple peaks, one of them (a pair), for example the image portion including the uppermost peaks 21L1 and 21R1 (for example, the image portion of about the upper 1/3 of the wire silhouette image 21 shown in Figure 2), is treated as the image portion to be processed for calculating the diameter d of the wire rope 1, which will be described below. This is to avoid the maximum values of the left and right radii being selected from different peaks when the wire silhouette image 21 contains multiple peaks (to avoid discrepancies between the calculation positions of the left radius and the right radius), which will be described below.

The image portion of the wire silhouette image 21 to be processed is used to calculate the left radius rL and the right radius rR of the wire rope 1 using the left edge position _xL , the right edge position _xR , the midpoint position _xM , and the inclination _θ of the center line C determined for each line image. With reference to FIG. 5, the left radius _rL is expressed by ( _xM - _xL ) cosθ, and the right radius _rR is expressed by ( _xR - _xM ) cosθ (step 50). Even if the wire rope 1 is obliquely captured by the _digital camera 11 and the wire silhouette image 21 is thus inclined from the vertical direction of the frame image 20, the left radius _rL and the right radius _rR can be calculated correctly. There is no need to pay much attention to the attitude of the digital camera 11 capturing the image of the wire rope 1, which is particularly convenient when measuring the diameter of the wire rope 1 in use at the installation site of an elevator or the like.

By carrying out the above-mentioned processing for each of the line images included in the processing target image portion of the wire silhouette image 21, a left radius _rL and a right radius _rR are found for each line image included in the processing target image portion of the wire silhouette image 21. The sum of the found maximum value of the left radius _rL and the found maximum value of the right radius _rR is set as the diameter d of the wire rope 1 represented by the wire silhouette image 21 included in the processing target frame image 20 (step 60).

In the case of a traveling wire rope 1, image data representing multiple frame images 20 captured at different locations is recorded in the storage device of the recording/analysis device 14. The above-mentioned processing is performed for each of the multiple frame images 20, thereby calculating the diameter d of multiple locations in the longitudinal direction of the wire rope 1.

Figure 6 is a graph showing the measurement results of the diameter d of the wire rope 1 calculated using multiple frame images 20, with the vertical axis showing the diameter d of the wire rope 1 calculated using the algorithm described above, and the horizontal axis showing the position of the wire rope 1 (here, as an example, the number of peaks appearing on the surface of the wire rope 1). Three very large peaks appear in the graph in Figure 6, but these are the diameter including the marking tape that was previously wrapped around the wire rope 1, and do not represent the diameter of the wire rope 1. The diameter d is measured for each frame image 20, so the diameter d of the wire rope 1 is measured at multiple points along the longitudinal direction.

When the diameter d was measured at multiple points along the length of the wire rope 1 using the algorithm described above, the average value was 10.34 mm. When the diameter of the same wire rope 1 was measured at multiple points using a vernier caliper, the average value was 10.31 mm. It can be seen that by using the algorithm described above, the diameter of the wire rope 1 can be measured with almost the same accuracy as when using a vernier caliper.

1. Wire rope
11 Camera
12 Lenses
13. Light Source
14 Recording/analysis equipment
20 Frame Images
21 Wire Silhouette Images

Claims (8)

a camera including an image sensor that captures an image of the filament so that the entire filament is captured in the width direction and outputs image data including the filament in a subject image;
an illumination device provided on the opposite side of the camera across the filament; and an image processing device that processes image data provided by the camera,
The image processing device,
an edge position determining means for determining both end edge positions of a silhouette image of the filament for each line image constituting a frame image represented by image data provided by the camera;
a center line determining means for determining a midpoint position equidistant from both end edge positions determined for each line image by the edge position determining means, and for determining a center line of the silhouette image of the filament included in the frame image using the determined midpoint position; and a diameter determining means for determining a length between both end edges of the silhouette image of the filament along a line segment perpendicular to the center line for each line image, using the both end edge positions, the midpoint position and the inclination of the center line ,
The diameter determination means is
a radius calculation means for calculating a left radius and a right radius of a silhouette image of the filament, using each of the end edge positions, the midpoint position, and the inclination of the center line;
Image diameter measurement system for the striatum. the diameter determination means determines the length between the two end edges as a sum of the maximum values of the left radius and the right radius calculated for each line image by the radius calculation means;
2. The system for imaging diameter measurement of a filament according to claim 1 . The edge position determining means
determining, as both end edge positions, the center positions of peaks represented by the amount of luminance change between horizontally adjacent pixels of the line image;
3. The image diameter measuring system for a striated body according to claim 1 or 2 . The edge position determining means
The positions showing the maximum value of the luminance change amount of the horizontally adjacent pixels of the line image are determined as the end edge positions.
3. The image diameter measuring system for a striated body according to claim 1 or 2 . the center line determined by the center line determination means is a regression line calculated by a least squares method using a midpoint position determined for each line image;
A system for measuring the diameter of a filament according to any one of claims 1 to 4 . The filament is a wire rope formed by twisting together a plurality of strands,
The diameter determination means is
Using an image portion including one peak of the silhouette image of the wire rope included in the frame image, a length between both end edges of the silhouette image of the filament is determined along a line segment perpendicular to the center line.
2. The system for imaging diameter measurement of a filament according to claim 1. an image data input means for receiving an input of image data representing a frame image including an entire subject image of the filament in the width direction;
edge position determining means for determining both end edge positions of the image of the filament for each line image constituting the frame image;
a center line determining means for determining a midpoint position equidistant from both end edge positions determined for each line image by the edge position determining means, and for determining a center line of the image of the filament included in the frame image using the midpoint position determined for each line image;
a diameter determining means for determining, for each line image, a length between both edge ends of the image of the filament along a line segment perpendicular to the center line, using the both edge positions, the midpoint position and the inclination of the center line ;
The diameter determination means is
a radius calculation means for calculating a left radius and a right radius of the image of the filament by using each of the end edge positions, the midpoint position, and the inclination of the center line;
Image diameter measurement device for the striatum. A program for measuring the diameter of a filament using image data representing a frame image including the entire filament in the width direction in a subject image,
determining edge positions at both ends of the image of the filament for each line image constituting the frame image;
determining a midpoint position equidistant from both end edge positions determined for each line image, and determining a center line of the image of the filament included in the frame image using the midpoint position determined for each line image;
determining, for each line image, a length between both edge ends of the image of the filament along a line segment perpendicular to the center line using the both edge positions, the midpoint position, and the inclination of the center line ;
In determining the length between both end edges of the image of the filament, a left radius and a right radius of the image of the filament are calculated using each of the end edge positions, the midpoint position, and the inclination of the center line;
Image diameter measurement program for the striatum.

Images