JPH0746382B2 - Image processing method - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing method in an apparatus for detecting edges of a detection target based on a luminance signal from an image detection camera.

[Conventional technology]

Generally, as a method of detecting the edge of the detection target from the luminance signal from the image detection camera in this type of device, the luminance signal from the image detection camera can be analog-digital converted and then reread immediately. The image is stored in a different memory, and various image processing such as noise removal and edge enhancement processing is performed on all the brightness data to detect the edge position of the detection target.

[Problems to be Solved by the Invention]

By the way, in the conventional method as described above, in order to detect the edge of the detection target, for a large amount of luminance data,
Since noise removal and edge enhancement processing and image processing were performed, there was a problem that a large amount of processing time was required when processed by software. For this reason, when the object is moving at a high speed, a method of configuring image processing by hardware (electronic circuit) has been generally used, but this method is used for changing / adding processing contents. It had a problem that it could not respond sufficiently.

The present invention intends to provide a new image processing method which solves the above-mentioned problems with respect to the image processing for a simple object such as a strip to be detected.

[Means for Solving the Problems]

Therefore, the image processing method of the present invention is applied to a device for detecting an edge of a detection target object based on a luminance signal from the image detection camera while the detection target object moves in a direction perpendicular to the scanning direction of the image detection camera. At the start of detection, for all the two-dimensional input image data, addition processing in the moving direction of the detected object is performed to convert the two-dimensional image data into one-dimensional data,
Next, differential processing is performed in the direction perpendicular to the moving direction to detect the edge of the detection target, and after detecting the edge, the moving speed of the edge is calculated from the edge position detected last time and the edge position detected two times before. Along with the calculation, the image capture period and the same speed are multiplied to obtain the movement amount, which is added to the previous edge position to obtain the estimated edge position of this time, and further,
The effective data width N set in advance based on the maximum value of the edge moving speed is added to or subtracted from the estimated edge position to determine the processing range this time, and the processing for obtaining the edge position is performed, so that the edge speed is increased. It is characterized in that it is possible to detect and follow the movement of the edge.

[Action]

Originally, noise removal, edge enhancement processing, and image processing are performed to detect edges of an object with high accuracy, but they are performed for all luminance data input from the image detection camera. There is no need to In particular, if the object has a simple shape such as a strip, it may be carried out only for data in a limited range near the edge.

There is a sufficient brightness difference between the detection target that moves in the direction perpendicular to the scanning direction of the image detection camera and its surroundings, and the amount of movement of the detection target in the camera scanning direction is sufficient compared to the detection width of the camera. If it is small, image processing is performed on all brightness data at the beginning of detection, and after detecting an edge, the edge position detected last time is multiplied by the change speed in the camera scanning direction at the same position by the sampling time. The position to which the value is added is regarded as the estimated edge position this time, and image processing is performed on the data for the effective data width N that can be changed by software, centered on the estimated edge position, and thus accurate The edge position can be detected in a short time, and the same effect as in the case where the image processing is performed on all the data can be obtained, and the software processing can realize a sufficiently high-speed processing.

As a result, by extracting an effective data portion from a large amount of digitally converted data of the luminance signal, it is possible to reduce the number of luminance data subject to the image processing performed after the extraction and even when the image processing is executed by software. , Sufficiently high image processing speed is obtained compared to the moving speed of the detection target,
In addition, it is possible to maintain the edge detection accuracy at the same level as in the case where image processing is performed on all conventional data.

〔Example〕

An image processing method according to an embodiment of the present invention will be described below with reference to FIG. 1. FIG. 1 shows the positional relationship between the image detection camera and the detection target (strip), and the scanning direction of the camera and the moving direction of the detection target. It shows the relationship. 1 in FIG. 1
Is a camera for image detection, 2 is a camera scanning direction, 3 is an object to be detected, and 4 is a moving direction of the object to be detected.

The luminance signal from the image detection camera 1 installed at the position as shown in FIG. 1 is converted from analog to digital, and the scanning direction of the camera 1 is plotted on the horizontal axis and the luminance is plotted on the vertical axis. Is. In FIG. 2, reference numeral 5 denotes the current edge position and the current estimated edge position obtained from the changing speed of the edge position in the camera scanning direction, and N is the effective data width set by software. The operation of the present invention will be described with reference to FIGS. 1 and 2 while comparing it with the prior art.
Usually, it is not clear where the steeply sloping portion (edge portion) of the luminance shown in the graph of FIG. 2 exists in the luminance signal from the camera. Therefore, conventionally, all luminance data has been the target of image processing. Therefore, if the data length of the luminance data is 8 bits and the resolution in the scanning direction is 512 points, assuming that the data of 10 lines is averaged for noise removal, 512 x 8 x 10 = 5120 bytes (1 byte = 8 bits ) Data must be processed, which is a factor that slows down the processing speed.

Therefore, in the present invention, since image processing is performed on all luminance data at the start of edge position detection, addition processing in the moving direction of the detected object is performed for all two-dimensional input image data to convert the two-dimensional image data into one-dimensional data. Then, the differential processing is performed in the direction perpendicular to the moving direction to detect the edge of the detection target. After the edge of the object is once detected in this way, the moving speed of the edge is calculated from the previously detected edge position and the edge position detected two times before, and the moving amount is calculated by multiplying the same by the image capturing cycle. , And this is added to the previous edge position to obtain the current estimated edge position.

Further, the effective data width N in which the maximum value of the edge moving speed is set in advance is added to or subtracted from the estimated edge position to determine the processing range this time, and the processing for obtaining the edge position is performed, so that the edge is quickly detected. It is possible to detect and follow the movement of the edge, and reduce the brightness data that is the target of image processing up to NX2 calculated by the moving speed of the detection target in the scanning direction. The speed can be obtained, and the edge detection accuracy can be maintained at the same accuracy as when the image processing is performed on the conventional all-luminance data.

〔The invention's effect〕

As described above, according to the image processing method of the present invention, the following effects can be obtained.

(1) In the method of the present invention, in order to detect the edge of the detection object that moves in the direction perpendicular to the scanning direction of the image detection camera, the edge position of the previous detection and the edge scanning position High-speed and high-accuracy edge detection by software is performed by using the corrected edge position as the estimated edge position this time and performing image processing only on the luminance data for the effective data width N centering on the estimated edge position. Made possible.

[Brief description of drawings]

FIG. 1 is a plan view showing a related arrangement state of an image detection camera and an object to be detected according to an embodiment of the present invention, and FIG. 2 shows an analog-digital conversion of a luminance signal from the image detection camera, It is a graph explanatory view showing the scanning direction of the camera on the horizontal axis and the luminance on the vertical axis. 1 ... Camera for image detection, 2 ... Camera scanning direction, 3 ...
… Detection object, 4 …… Detection object moving direction, 5 …… Estimated edge position, N …… Effective data width.

Claims (1)

[Claims] 1. An apparatus for detecting an edge of a detection target based on a luminance signal from the image detection camera while the detection target moves in a direction perpendicular to a scanning direction of the image detection camera. Sometimes, the two-dimensional input data is added to the moving direction of the detected object to convert the two-dimensional image data into one-dimensional data, and then the differential processing is performed in the direction perpendicular to the moving direction. Edge detection of the detection target is performed, and after detecting the edge, the moving speed of the edge is calculated from the previously detected edge position and the edge position detected two times before, and the moving speed is multiplied by the same speed as the image capture cycle. The amount is calculated, this is added to the previous edge position to obtain the estimated edge position this time, and the effective data width N preset based on the maximum value of the edge moving speed is added to or subtracted from the estimated edge position to obtain the current value. Determines the processing range by performing the process for determining the edge position, an image processing method is characterized in that it possible to detect that follows the movement of the edge detects the edge at a high speed.