JPS60214084A - Picture data processing system - Google Patents

Abstract

PURPOSE:To detect vertexes with a high reliability in a short processing time by scanning only boundaries of an object in the eight-nearby point scanning system. CONSTITUTION:The position of scanning start position AS is set to coordinate positions X and Y in a memory. At this time, an initial value is stored in the position, where a parameter of the preceding tracing processing should be stored, of the memory. The tracing processing of nearby boundaries is performed, and it is checked whether a position obtained till then on the memory is maximum in the (y) direction or not. If it is maximum, a value YMAX on the memory is updated. If it is not maximum, tracing is terminated if tracing of boundaries goes beyond a frame in the (x) direction, and the control is returned to the tracing processing otherwise. When tracing is terminated, the value YMAX in the memory indicates an obtained vertex.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a vertex detection power formula for an object in single image data processing.

[Prior art]

FIG. 1 is a block diagram of a data processing device that detects the tip of an object using a conventional sword method. Imaged by camera 1 f'
Lfc A video signal of an object (not shown) is converted into digital data by an input circuit 2.

The coordinates corresponding to the X and Y coordinates of the image of the object are stored in the memory 3 in an arrangement as shown in FIG. The processor 4 consists of a read-only memory and a program? A process of reading out the data of the mesori 3 according to the program read out from the stored memory 5 and the memory 6 consisting of a random access memory, and detecting the tip of the object according to the flowchart shown in FIG. do.

FIG. 2 shows the arrangement of data in the memory 3 inside the memory mange. ASσ memory 30 indicates the starting position of travel strain, and memory 3 is on the sequential y-axis in the direction of the arrow on the X-axis? Scanned. A
The point is yF, the vertex of the target object.

A series of processing by the processor 4 will be briefly described in accordance with the 'fRf'L diagram shown in FIG. Here, to the vertex is y
Place the point on the axis that indicates the maximum value. When the start position As is set at step S1, the set meal starts in the y direction as indicated by the arrow at step S2.

Read the contents of memory 3. In step S3, it is determined whether there is a change in the contents of the memory 3 (for example, whether there is a change from l to 0), that is, whether there is a boundary, and in step S4, it is determined whether or not it is the final y address. If the judgment is negative, the process returns to step 8.2, and if the scanning is completed, the process returns to step S.
In Step 5, the current maximum value is set as the vertex A, and in Step S6, if it is the highest P:x address, the process ends, and if it is, the process returns to Step S2. If the result in step S6 is negative, the process proceeds to step S7, where the boundary value is compared with the current maximum value, and if the value is larger than n', the process proceeds to step S8.
What is the current maximum value? This will update.

Finally, the current maximum value of f'Lyc is the vertex A.

In the conventional method, the number of times the image memory is accessed is large (1 or more), so processing time is precious, and the number of times the image memory is accessed is high. Since the vertex is detected by detecting the vertex, if there is erroneous data with a large value (for example, the three points shown in FIG. 3), this method has the disadvantage that the vertex may be incorrectly determined.

[Summary of the invention]

This invention was made in order to eliminate the drawbacks of the conventional methods as described above, and by scanning only the boundaries of objects using the 8-neighborhood method, it is possible to achieve high reliability with a very short processing time. An image data processing method that can detect high vertices?
What to offer? The purpose is

[Embodiments of the invention]

Hereinafter, an embodiment of the present invention will be described. Fifth
The figure is a flowchart showing a series of processes according to the method of this invention. A data processing device that performs such processing is shown in No. 1-, and is controlled by a processor 4. memory 3
The outline of scanning is as shown by the arrow in the memory map diagram of FIG.
s and proceed along the boundary.

In FIG. 5, in step P1, parameters CX, X and Y are stored in memory 3 by the program.
v Load from memory 3. Here, X and Y are the coordinate positions of the target pixels I and J in the memory 3 as shown in FIG. show. 'F7C and ICX were the pixel works of interest in the previous boundary tracking process.

This is a parameter indicating the coordinate position of J. Step P2~
At P8, the scanning start position within 8 neighborhoods (field P indicated by 1 in FIG. 10) r
) Branches to the begging process. The processing in steps P9 to P16 is a calculation process for providing the positions on the memory 3 of eight neighbors around the pixel of interest. Are XT and YT near 8? This is a parameter indicating the position of the memory 3, and is stored in the memory 6.

In steps P17 to P24, whether the contents in the memory 3 indicated by XT and YT given in steps P9 to P16 are 0 or not? To check. if.

If it is O, move the contents of XT and YT in memory 6 to the place in memory 6 where the values of X and Y are arranged,
Update the contents of ICX in memory 6. If it is not O, move on to the process of calculating the other position within the 8 neighborhood.

FIG. 10 is given by the value of ICX.

It shows the scanning start position and scanning order within the 8-neighborhood.

In the figure, N is the vicinity of the pixel of interest given by the X and Y parameters in Fig. 5, 0 indicates the pixel of interest in the previous process, and the numbers 1゜2 to 7 in the block are the scanning order. It shows.

FIG. 9 is a flowchart of the processing according to the method of this invention. In step S11, the position of the disclosure position AS shown in FIG.
Cent is placed at the coordinate positions X and Y in the memory 6 used in FIG. At this time, the initial value is stored in the storage location of the parameter ICX in the memory 6.

In step S12, the 8-neighbor boundary tracking process shown in FIG. do. In step S13, it is checked whether the position on the memory 3, which has been nl xl in step S12, is the maximum (MAX) for y1 times. if.

At MAX, an proceeds to step 814 and YM in memory 6 is
The value of AX is updated, and if n is not MAX, the process moves to step 815. In step s15, if the boundary tracking exceeds the frame in the X direction, the tracking ends, and if not, the process returns to step 812.

When the tracking ends, YMAX in the memory 6 is the vertex A that has been reached.

〔Effect of the invention〕

As described above, according to the present invention, since only the boundaries of objects are fixed, the number of accesses to the memory for one image is reduced, processing time is shortened, and 1-fC noise is not affected. However, it has the effect of detecting highly reliable vertices only in noise that is in contact with objects.

[Brief explanation of the drawing]

Figure 1 is a block diagram of the data processing device, Figures 2 and 3
The figure is a memory map diagram of the memory shown in FIG. 1, and FIG. 4 is a flowchart of a conventional image data processing power formula. FIG. 5 is a flowchart of processing according to an embodiment of the present invention. Fig. 6 is a memory map diagram of the memory showing the set meal direction, Fig. 7 is a ``f coordinate diagram showing the 8 neighborhood boundaries, Fig. 8 is a memory map diagram of the mesori that shows the 71! is an outline flow n diagram according to the present invention, and FIG. 10 is a diagram showing the order of set meals in 8 neighborhoods. 1...Camera, 2...Input circuit, 3, 5.6...
Memory, 4... Processor. In Figure 1, the same reference numerals indicate the same parts. Patent applicant Mitsubishi Electric Corporation (2 others) -:,'p Figure 3 Figure 6 Figure 1 Figure 8 Figure 9 Figure 10 ICX=OICX=I ICX=3 KX=4 ICX=6 ICX=7 ICX=2 ICX=5

Claims (1)

[Claims] Vertex Y to be detected! The pixel data of the object is arranged and stored in a memory corresponding to the observed shape of the object, and when the boundary of the object is detected from the data read from the memory, the data corresponding to 8 neighboring points of the boundary is stored. An image data processing method that detects the vertex by reading from the memory, detecting a boundary at an adjacent memory location, and determining whether the boundary gives a maximum value on coordinates.