JPH0663752B2 - Recognition method for linear objects - Google Patents

Description

Detailed Description of the Invention [Industrial applications]

The present invention relates to a method for recognizing a linear object, and more particularly to a method for recognizing the shape of a linear object such as a wire stretched in a space.

[Prior art]

FIG. 11 is a block diagram for explaining the conventional position measuring method of a linear object described in, for example, Proceedings G-78 of the Western Conference of the Electrical Association of Japan in 1986. 1 is an objective lens,
Reference numeral 3 denotes an image pickup device (camera) that converts an image formed by the objective lens 1 into an electric signal, 4 an image processing device that receives a video signal supplied from the image pickup device 3 and performs image processing, and 5 a visual field of the objective lens 1. A linear object as an object to be measured located inside is a moving table 6 for moving the linear object 5. Next, the operation will be described. An image of a linear object 5 as an object to be measured placed on the movable table 6 is formed on the image pickup surface of the image pickup device 3 by the objective lens 1. Here, since the linear object 5 has a three-dimensionally changing shape, the linear object 5 passes through the imaging planes of the objective lens 1 and the image sensor 3 and the focal plane 100 determined by the focal length of the objective lens 1. Only the part is imaged with good contrast, and the image of the other part is blurred. On the other hand, when the movable table 6 is moved up and down, the linear object 5 passing through the focal plane 100 is moved along with it.
The position of changes. On the other hand, the video signal output from the image sensor 3 is subjected to video processing in the image processing device 4 to extract the image of the linear object 5 from the background, thereby extracting the contrast between the object and the background. Then, from the image data processed in this way, the position of the moving table 6 having the maximum contrast is obtained and stored for each position on the imaging surface (called a pixel when digital image processing is performed). By repeatedly performing such processing while changing the moving table 6 by a small amount, the position of the moving table 6 having the maximum contrast is obtained and stored. When the processing in the moving range of the moving table 6 is completed, the stored value of the position of the moving table 6 that maximizes the contrast represents the height of the linear object 5.

[Problems to be Solved by the Invention]

The conventional linear object shape measuring method requires a long time for measurement because it is necessary to displace the moving table 6 by a small amount and to capture and process an image each time. Also, the moving table 6
Since the object is imaged and processed each time by displacing the object by a small amount, there is a problem that the linear object 5 whose shape changes with time or moving cannot be recognized. . The present invention has been made to solve the above problems, and an object thereof is to obtain a linear object recognition method capable of detecting the shape of a linear object with a single image capture. .

[Means for Solving the Problems]

A method for detecting a linear object according to the present invention uses a plurality of image pickup devices having different focal planes to simultaneously photograph the same field of view, evaluates the contrast of the linear object image for each image, and determines the line width of the linear object. Recognize and measure the position where each focal plane intersects the linear object, and if it is not the intersection, measure the amount of deviation from the focal plane, and combine the above various quantities for each focal plane to determine the shape of the linear object. It is to be calculated.

[Action]

The method for detecting a linear object according to the present invention forms the focal planes of different heights by the optical path splitting means and the plurality of image pickup elements, and makes it possible to observe the scene for each focal plane at the same time,
In the image processing device, the position, angle, and the amount of deviation from the focal plane where the linear object intersects each of the focal planes are measured, and the intersection position of each surface is interpolated using the angular displacement amount. Therefore, it is possible to recognize the entire shape by capturing the image once.

Examples of the invention

An embodiment of the present invention will be described below with reference to the drawings. First
In FIG. 2 and FIG. 2, 1 is an objective lens, 21, 22, 23 are half mirrors mounted on the optical axis of the objective lens 1,
Reference numerals 31,32,33 denote image pickup elements arranged on the optical axis divided by the half mirrors 21,22,23, and 4 denote the image pickup elements 31,32,
33 is an image processing device that processes a video signal output from 33 to calculate the shape of a linear object. Reference numeral 5 is a linear object as a recognition object. Next, the operation will be described. The image of the linear object 5 is the objective lens 1
After being divided by the half mirrors 21, 22, and 23, an image is formed on the image pickup surface of each of the image pickup devices 31 to 33. Here, the optical distance between the objective lens 1 and each of the image pickup devices 31, 32, 33 is different from each other, and when the object is on the plane 101, the object is focused on the image pickup device 31 and is on the plane 102. Sometimes the image is focused on the image sensor 32, and when the object is on the plane 103, the image is formed on the image sensor 33. Further, assuming that the linear object 5 is arranged so as to intersect with the above-mentioned planes 101 to 103, the line width is thinnest and the image contrast is maximized at the portion intersecting with each of the planes. The object images shown in FIGS. 3 to 5 are captured. Image processing device 4
In FIG. 3 supplied from each of the image pickup devices 31 to 33.
An image obtained by shooting the object image shown in FIG. 5 is loaded into the frame memory, and image processing is executed according to the flowchart shown in FIG. That takes the image signal supplied from the imaging elements 31 to 33 in the frame memory in step ST _1, and sets to zero the image signal in step ST _2. Next, in the step ST _3, searches the linear object image by processing the image data stored in the frame memory, obtains the center line. For example, when illuminating a linear object from above or from below,
The linear object is captured as a dark image as compared with the surroundings, as shown in FIGS. Therefore, in the density change pattern shown in FIG. 7, if the minimum value of the brightness in the image is traced, the center 701 of the linear object 5 can be obtained. Next, in step ST _4, determine the concentration change pattern such as Figure 7 to measure the direction density perpendicular to the centerline. Then, the width 702 and the contrast 703 of the linear object image are obtained from this density change pattern. Since these values are determined by the amount of deviation from the focal plane of the object,
If the actual line width of the linear object is known and constant, on the contrary, the amount of deviation can be known from the image width 702. Then step S
At T ₅ , the crossing position of the focal plane and the linear object 5 is
It is obtained by searching for a point where the width of the linear object image matches the actual line width of the linear object and gives the maximum contrast. Next, in step ST _6, obtains the amount of deviation from the focal plane at the intersection location and each point of each focal plane. In step ST _7, in accordance processing screen makes a determination is made as to whether the last screen, if the determination result is no, the flow returns to step ST ₃ after incrementing the frame number in step ST ₈ Let the process run. By performing such a determination, the series of processes described above is performed on each screen of the image pickup elements 101 to 103, and the intersection position for each focal plane and the deviation amount from the focal plane of the linear object in the vicinity of each point are calculated. Ask. Next, the process for obtaining the amount of deviation from the focal plane at the intersection location and each point of each focal plane reaches the last frame, step ST ₉ determination in step ST ₇ becomes YES
Move to. In step ST _9,. 31 to the image sensor
After correcting the difference in magnification and the deviation of the central axis for each 33, the screens are overlapped to determine the shape of the linear object. That is, an object of a known shape is separately photographed by the image pickup devices 31, 32, and 33 in advance, the shift, enlargement, and reduction ratios of each image are measured, and this value is used to determine the magnification and center between the image pickup devices. the process proceeds to step ST _{1 0} the deviation is corrected. In step ST _{1 0,} determines the position that intersects the respective focal plane, further by the amount of deviation from the focal plane, performs the operation to calculate the passing position of the portion between the focal plane succeed between intersections And it ends. By performing such processing, it becomes possible to easily recognize the shape of the entire linear object. Although no special illumination is used in the above embodiment, a half mirror 20 is provided on the optical axis of the objective lens 1 as shown in FIG. 8 to collect the light emitted from the illumination light source 70. After collecting with the lens 71, the coaxial illumination for illuminating the linear object 5 by supplying to the half mirror 20, or after collecting the light emitted from the illumination light source 70 with the condenser lens 71 as shown in FIG. By using transmissive illumination that illuminates the linear object 5 from the lower side via the diffusion plate 73, it becomes easy to extract a change in contrast. When the background is dark and transmitted illumination is difficult, the illumination light source 70 is arranged at a target portion with respect to the optical axis of the objective lens 1 as shown in FIG. By using diffused illumination in which the linear light 5 is applied to the linear object 5 after being diffused by the diffuser plate 73, it becomes easy to extract the linear object 5 brightly and the background dark. Further, in the above embodiment, the half mirror is used for dividing the optical path, but a means such as a prism may be used, and the dividing method is also divided into two and then divided into two.
The same effect can be obtained by using means such as a split optical system. Furthermore, in order to improve the accuracy in the depth direction, the number of image pickup elements is increased, or the interval between the focusing planes 101 to 103 is made variable,
It may be configured to recognize in two stages.

【The invention's effect】

As described above, according to the present invention, since images on a plurality of focal planes are simultaneously captured by one image capturing process, it is possible to capture images while changing the distance between the object and the imaging device. There is an effect that it becomes unnecessary, the processing speed is increased accordingly, and the shape of a linear object whose shape changes temporally or is moving can be easily recognized.

[Brief description of drawings]

1 and 2 are a perspective view and a side view showing a configuration for explaining a method of recognizing a linear object according to an embodiment of the present invention, and FIGS. 3 to 5 are respective photographings shown in FIG. FIG. 6 is a flow chart showing a method for recognizing a linear object, FIG. 7 is a density change pattern diagram showing the relationship between the line width of the linear object and the contrast, and FIG.
FIG. 10 is a block diagram for explaining another embodiment of the linear object recognition method according to the present invention, and FIG. 11 is a block diagram for explaining a conventional linear object recognition method. Reference numeral 1 is an objective lens, 3,31 to 33 are image pickup elements, 4 is an image processing device, 5 is a linear object, and 100 to 103 are focusing planes.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Kazuo Aida 8-1-1 Tsukaguchi Honcho, Amagasaki City, Hyogo Sanryo Electric Co., Ltd. Communication Equipment Factory (72) Masahiro Sasakura 8 Tsukaguchi Honcho, Amagasaki City, Hyogo Prefecture No. 1-1 Sanryo Electric Co., Ltd. Communication Machinery Works (72) Inventor Hiroshi Ozaki 8-1-1 Tsukaguchihonmachi, Amagasaki City, Hyogo Prefecture Sanryo Electric Co., Ltd. Communication Machinery Works (56) Reference Japanese Patent Laid-Open No. 61 -169705 (JP, A)

Claims (1)

[Claims] 1. A method for recognizing the shape of a linear object arranged in a space, wherein light passing through an objective lens arranged so as to face the linear object is divided into a plurality of pieces, and the light is divided into a plurality of pieces. Contrast and image thickness are measured by providing image sensors at positions where the distance from the objective lens is different for each optical axis, and subjecting the linear object image captured by each image sensor to image processing by an image processing device. By determining the position where the linear object intersects with the focusing plane determined for each image sensor and the distance to the focusing plane, the linear object image captured by each image sensor is determined. The shape of the linear object arranged in the space is recognized by connecting the position where the linear object intersects the focusing plane and the distance to the focusing plane. Recognition method Jo object.