JPH02127393A - Object position detecting device - Google Patents

Abstract

PURPOSE:To automatically detect the upper face position of a container in an on-line operation by processing the image data photographed by a TV camera with a binary processing device, a contour line detecting device, a straightening device and an arithmetic device in a detecting device for positioning a container. CONSTITUTION:An A-D converter 2 digitizing the image signal photographed by a TV camera 1 with 8 bits and an image memory device 3 are provided. The image data of a subject container are: 1) converted into a binary image by a binary processing device 4 to extract the container. 2) The contour line coordinates of the container on the binary image are detected by a contour line detecting device 5. 3) The contour line is straightened by a straightening device 6 and an equation is obtained. 4) The center of gravity and inclination of the container shown by four apexes are determined by an arithmetic device 7. A display device for the figure position determined by the arithmetic device 7 is provided. The automatic operation of a container crane or the like can be performed.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an object position detection device that is applied to positioning an object (container) in an automatic operation device such as a transformer 7-arc crane or a container crane.

[Conventional technology]

Conventionally, operating a transfer crane or a container clay is a manual operation performed by an operator, and the position of a container is also detected visually by the operator.

[Problem to be solved by the invention]

As mentioned above, conventional crane operation work was manual work performed by the operator, and container position detection was performed visually, so it was not possible to automate the movement work of containers and the like.

The object of the present invention is to detect the position of a container, etc., which is necessary for rounding to automate the movement of objects such as containers of transfer cranes and container cranes, which was previously performed manually by operators. The purpose is to provide equipment.

[Means to solve the problem]

The object position detection device according to the present invention includes a TV camera 2, an A-D converter that digitizes a video signal captured by the TV camera in 8 bits, and video signal data (image data) digitized in 8 bits. an image storage device that stores
a binary processing device that converts the 8-bit image data into a binary image; a contour detection device that detects the coordinates of the contour of a figure on the binary image; and a contour detection device that straightens the contour and calculates the equation of the straight line. A linearization device for determining the position of the target figure based on the equation of the straight line determined by the linearization device, and a display device for displaying the position of the figure determined by the calculation device. It is characterized by becoming.

[Effect]

According to the present invention, the position of a target object (container) is detected by (1) binarizing the image data of the target container; (2) extraction of the container; and (2) contour coordinates of the container. (3) Linearization of the contour line, and (4) Calculation of the center of gravity and inclination of the container.The operation of the above processing will be explained below.

(1) Extraction of containers by binarization processing The video signal of the TV camera that imaged the container is converted into an 8-bit image r-ta by an A-D converter, and then converted into an 8-bit image r-ta by a binarization processing device. The container can be extracted by converting it to a value image. FIG. 2 shows a processing flow diagram of the binarization processing device.

A threshold k for binarization is calculated for 8-bit image data using Otsu's method. (References for Otsu's method: Nobuyuki Otsu, Automatic threshold selection method based on discriminant and least squares criterion, IEICE Journal 8014 Vol. J63-DA4, P.
, 349-P, 356). Otsu's method is a method in which a density histogram of 8-bit image data is created, and when the histogram is divided into two classes at a certain density, the density k that maximizes the interclass variance σ, 3 is used as a threshold. be.

The 8-bit image data is binarized using the determined threshold value k. (The area corresponding to the container is 1°, and the other areas are 0.) Count the number of pixels in the area corresponding to the container for the binarized image data, and combine the result with the expected number of pixels occupied by the container ( The expected number of pixels for the container is
Determined by the size of the container, the viewing angle of the camera, and the distance between the camera and the container. ), and if the difference is within a certain tolerance range, the binarization process ends.

If not, perform an AND operation between the binarized image and the original 8-bit image, and then compute the threshold value k for binarization using Otsu's method on the resulting image. Repeat.

In short, each time the binarization processing is repeated, the binarization processing device limits the target processing area, and when the processing area is different from the expected area (number of pixels) of the container by K, the binarization processing device After completing the processing, the container is extracted by regarding the area as a container.

(2) Detection and extraction of container contour coordinates The contour coordinates of the nine containers (X-axis: horizontal axis of the camera, Y-axis: vertical axis of the camera) are determined by a contour detection device. Third
The figure shows a processing flow diagram of the contour line detection device.

The t2 binary image data is arrayed IP (M, N) (M,
N is the number of pixels in the horizontal and vertical directions, respectively), then (
(See FIG. 4 (2)), the contour points on the binary image data can be found, for example, by using the origin as the starting point and scanning in the lower right direction to find a point with 1" wobbling.

Let the coordinates of that point be X(1) and Y(1).

To search for contour points, for example, ■ the contour of the container must be closed, ■ search in a clockwise direction (counterclockwise direction is also possible)
Assuming that, the current contour point P is (X(10, Y(K)),
The previous contour point q is (X(K-1), Y(K-1))
, the next contour point R is (X(K+1), Y(K+1)), then for the data in the 3×3 neighborhood centered on the current contour point P, start from the previous contour point Q and move to the right. There is a method of checking the data of the 3×3 neighboring coordinates in the rotational direction (counterclockwise direction) and setting the coordinate where @1" is first set as the next contour point (R) in FIG. 4 CB.

The termination condition for contour point search is the coordinates of the next contour point (X(K
+1 ), Y(K+1) ) and the initial point (X(1)
, Y(1) >2>E - This can be achieved by confirming that.

(3) Straightening the contour line Although the contour line of a container should originally be a straight line, there are many cases where this is not the case due to external disturbances. Therefore,
A linearization device linearizes the contour line based on the contour point data using the least squares method or Hough transformation, and calculates its equation.

(4) Calculation of the center of gravity and inclination of the container Since the top surface of the container is rectangular, four straight line equations can be obtained using the straightening device. The arithmetic unit corrects the slopes and y-intercept values of these four straight line equations based on the fact that the top surface of the container is rectangular. Furthermore, the intersection points of the four corrected straight line equations are found, and these points are set as the vertices of the container (rectangle). (4 points are obtained.), the vertices of these four containers (rectangles), the center of gravity and the inclination of the containers (rectangles) (
) for the horizontal coordinate axis.

As described above, the container position (center of gravity and inclination) can be detected based on the flow.

〔Example〕

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention. The structure and operation of an embodiment of the present invention will be explained below based on FIG.

1 is a TV camera, 2 is an A-D converter, and 3 is an image storage device. A TV camera 1 is installed vertically downward and images the top surface of a target container. The imaged Eirin signal is A-
The data is converted into 8-bit digital data (image data) by the D converter 2 and temporarily stored in the image storage device 3.

4 is a binarization processing device, 5 is a contour detection device, 6 is a linearization device, and 7 is an arithmetic device.

The 8-bit image data read from the image storage device 3 is binarized by the binarization processing device 4, and containers are extracted. Next, the contour line detection device 5 calculates the contour line of the container on the image from this binarized data. Furthermore, the linearization device 6 removes disturbances (noise) included in the contour line, straightens the contour line, and obtains its equations (four equations are obtained). The arithmetic unit 7 corrects the slope and y-intercept of the equation of the straight line based on the fact that the top surface of the container is rectangular, and finds the intersection of the corrected straight lines and uses it as the coordinates of the vertex of the container (rectangle). The coordinates of this vertex are used to calculate the center of gravity and inclination of the top surface of the container, and the display device 8
to be displayed.

As described above, the device of the present invention makes it possible to detect the center of gravity and inclination of the top surface of a container.
Automated operation (automatic positioning) can be realized.

〔Effect of the invention〕

According to the present invention, the detection of the position (center of gravity and inclination) of the top surface of a container, which was conventionally performed visually by an operator, can be performed online and automatically. Furthermore, by applying the device of the present invention, there is a ripple effect in that the operation of transfer cranes, container cranes, etc. can be automated.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention;
FIG. 2 is a diagram showing the processing flow of the binarization processing device in FIG. 1, FIG. 3 is a diagram showing the processing flow of the contour line detection device in FIG. 1, and FIG. FIG. 4B, which is an explanatory diagram of the binary image data IP (M, N), is a principle diagram of contour point detection in FIG. 1. 1...TV camera, 2...A-D converter, 3...
Image storage device, 4... Binarization processing device, 5... Contour detection device, 6... Straightening device, 7... Arithmetic device,
8...Display device.

Claims (1)

[Claims] A TV camera and the video signal captured by this TV camera are
An A-D converter that digitizes in bits, an image storage device that stores 8-bit digitized video signal data (image data), and a binarization processing device that converts the 8-bit image data into a binary image. a contour line detection device that detects the coordinates of the contour line of a figure on a binary image; a linearization device that straightens the contour line and obtains the equation of the straight line; What is claimed is: 1. An object position detection device comprising: an arithmetic device for determining the position of a target figure; and a display device for displaying the position of the figure determined by the arithmetic device.