JPS5927374A - Boundary line tracking circuit - Google Patents

Abstract

PURPOSE:To track a boundary line at a high speed, by detecting the following boundary point from a data of eight picture element being adjacent to a present point which is tracking a boundary line, and also obtaining an output of its moving direction vector. CONSTITUTION:A picture of an object to be measured is inputted by a television camera 1, is binary-coded, and is stored as a picture data in a picture memory 3. A position of one point on a boundary line of the object is detected by a start point detecting circuit 4 from said data. When the position of its point is provided as a present point to a block extracting circuit 5, a data of the adjacent block point is extracted at the present point. By providing this data to a tracking vector circuit 6, a boundary point is selected from eight points being adjacent to the present point, and also a moving direction vector is obtained. The position data of the boundary point selected by the position data of the present point and the moving direction vector is calculated, the boundary line is tracked by setting the selected boundary point as the present point, and a special feature of a shape of an object is detected.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention tracks boundaries of binarized image data stored in an image memory, and detects geometrical feature points of an object, for example, as shown in FIG.
The present invention relates to a circuit for tracking a boundary line of a binary image in a device for detecting the position of a corner point Pk, etc. as shown in b).

Configuration of conventional example and its problems In general, boundary line tracing and feature point detection of a binary image are performed as follows, and the procedure will be explained using FIGS. 1(2L) and (b).

1. Give a certain point PO on the object that forms the boundary between the object and the background as the starting point.

2. Assuming the given starting point P, ) as the current point PO, Figure 1 (
8 pixels 1 to 8 adjacent to the current point PO shown in a)
Based on the information regarding brightness, it is determined whether each pixel is inside or outside the boundary, and from the adjacent 8 pixels 1 to 8, pixels on the boundary line, that is, points on the object that form the boundary between the object and the background. Select P1.

3. Determine whether or not to continue tracing the boundary line, and if so, 2. 2. Set the point P1 selected in as the current point.
The following processing is performed to detect P2 to Pn.

4. The movement direction code that is sequentially output during the process of executing boundary line tracing, that is, the time series pattern of the differences of numbers 1 to 8 assigned to the adjacent eight pixels shown in FIG. Detect geometric feature points such as Pk.

Conventionally, such processing used a microcomputer,
Read data from image memory and save the data to C
This is done in software by performing arithmetic processing using the PU. However, in this kind of software processing, it takes several cycles to access data from memory via the CPU, and the target boundary point is determined by determining whether the read data is O" or 1". The process of detecting and calculating the position also requires operations such as bit testing and addition, and each step requires several tens of microseconds of processing time.From a practical standpoint, even faster processing is required. If required, it is difficult to implement using software processing.

OBJECTS OF THE INVENTION In view of the above problems, an object of the present invention is to speed up boundary point tracking.

Structure of the Invention The present invention is as described above.2. In this process, by giving information on the brightness (o") and dark (1") of the 8 pixels adjacent to the current point, one point on the boundary line is selected from among the 8 adjacent pixels, and its direction vector is determined. This is to speed up boundary line tracking.

DESCRIPTION OF EMBODIMENTS An embodiment of the present invention will be described below with reference to FIG. 2 and subsequent figures. First, in order to clarify the function of the boundary line tracing circuit, the overall configuration of the feature point detection device will be explained using FIG. 2.

A television camera 1 captures an image of the object to be measured,
The video signal is binarized by the binarization circuit 2 and the signal is stored in the image memory 3 as image data. From this image data, a start point detection circuit 4 detects the position of a certain point on the boundary of the object. By supplying the position data of the detected boundary point to the block extraction circuit 6 as the current point, image data of the block point adjacent to the current point is extracted. By supplying this image data to the tracking vector detection circuit 6, a boundary point is selected from among the eight points adjacent to the current point, and the moving direction vector from the current point to the selected boundary point is determined by the extraction direction indicating circuit 7. can get. The position data of the selected boundary point is calculated from the position data of the current point and the movement direction vector, and the same process is performed using the selected boundary point as the current point, and the boundary line of the object on the binary image memory is Track. The feature point detection circuit 8 detects the shape feature points of the object by detecting a specific pattern from time-series changes in the moving direction code sequentially output from the boundary line tracking circuit.

Next, a method for selecting a boundary point of an object from among points adjacent to the current point in the boundary line tracing circuit will be explained with reference to FIGS. 3 and 4. Here, (•) in the figure is a point within the boundary of the object to be measured and is stored as "1" in the image memory, and (○) is a point outside the boundary, that is, in the background, and is stored as "0" in the image memory. First, in Fig. 3, the data is set to “1” starting from the point adjacent to the current point P.
or “0”, that is, whether it is inside or outside the boundary.

If the data is "1", that is, inside the boundary, immediately check whether the data at the next point is "1" or 0.In this way, when "0", that is, a point outside the boundary is found, Then, the point that was 1'', that is, inside the boundary until just before that point, is determined to be the point on the boundary line 2'' between the object and the background, and is set as the next current point. In the above processing, first 0".

If the point at which the determination of "1" is made, that is, the point 1 in FIG. 3, is already 0", the point 8 is "0".

If the value is "1", point 8 is selected as the boundary point, and if the value is "0", it is determined that it is impossible to trace. It is assumed that even if all data from 1 to 8 are "1", it is still impossible to trace. The above processing is performed by the tracking vector detection circuit 6. In addition, the instructions for moving the extraction block from the current point to the selected boundary point are downward if the selected point is 1.2.3, 3.
If the angle is 4.5, move to the left, 6, 6. ．． This can be obtained by moving the extraction block upward by one pixel in the case of 7, and to the right in the cases of 7 and 8.1.

This process is performed in the extraction direction indicating circuit 7. Fourth
In the figure, 2 is selected as the boundary point for (al), 3 is selected for +b, and 4 is selected for (C) as the boundary point, and the extraction block moves downward, downward to the left, and to the left, respectively.

The circuit configuration for performing the above processing is shown in FIGS. FIG. 6 shows the tracking vector detection circuit 6, and FIG. 6 shows the extraction direction instruction circuit 7. When AND circuits 1 to 7 are connected as shown in Figure 6, in A-H, the circuit where "0" is added first in the inputs of ■ to ■ becomes "1", and the subsequent circuits become "○".
”.For example, 1 adjacent to the current point in Figure 3
When the image data of ~8 is added to ■~■ in Figure 5, it becomes 1'' in ■~■ and ``0'' first in ■, so A~C'=
If it is 1, it will be "1", and if it is next to D, it will be "0". Further A-H
Exclusive OR circuit 8 ~
By giving it to 14, the input becomes “1” in ■～■
, only the output of the exclusive OR circuit 1o that is “○” is “1”
In this case, an output in which the other values are "0" is obtained. However, as mentioned above, in ■, the A N D circuit 16 converts 6o in ■.
"1" is output when "1" is obtained in "1". Exclusive OR circuits 8 to 14 output "1"
, or the AND circuit 16 inputs ``1'' to ``1'' in two adjacent inputs that first changed to ``1'' - 0.
, i.e., let ■～■ correspond to the input that is
Binary codes of direction codes 1 to 8 corresponding to (1) and assigned to (1) to (2) are obtained from the outputs BO to B3 of the encoder 16. That is, the next boundary point to be the current point is selected from 1 to 8, and the encoder 16 obtains its binary code. However, if all of ■ to ■ are "o", it is impossible to trace, so the AND circuit 17 outputs 1 as an error.
” is output.Then, the three-person OR of ■～■ is shown in Figure 6.
By inputting the signals to the circuits 18 to 21, outputs in the moving direction from the current point to the selected boundary point, that is, in the vertical and horizontal moving directions, are obtained from U, D, L, and R, respectively. In the example of FIG. 6, since L is "1", the lower left direction, that is, point 3 among adjacent points 1 to 8, is set as the next current point.

Effects of the Invention As shown in Section 2, the boundary line tracing circuit of the present invention detects the next boundary point from the data of 8 pixels adjacent to the current point where the boundary line is being traced, and also detects its movement direction vector. By obtaining the output, processing that was conventionally performed by a CPU can be replaced with hardware, and high-speed processing becomes possible. ,

[Brief explanation of the drawing]

FIGS. 1(2L) and (b) are explanatory diagrams of boundary line tracing and feature point detection, and FIG. 2 is a block diagram of a geometric feature detection device in a boundary line tracing circuit according to an embodiment of the present invention. Figures 3, 4 (a), (b), and (C) are explanatory diagrams of the boundary line tracing process, Figure 5 is a circuit diagram of the tracking vector detection circuit in the present invention, and Figure 6 is the extraction direction in the present invention. FIG. 3 is a circuit diagram of an instruction circuit. 1... TV camera, 2... Binarization circuit, 3... Image memory, 4... Start point detection circuit, 6...・Block extraction circuit, 6・
. . . Tracking vector detection circuit, 7 . . . Extraction direction instruction circuit, 8 . . . Feature point detection circuit. Name of agent: Patent attorney Toshio Nakao and 1 other person No. 4
Figure (α) (C>

Claims (1)

[Claims] A storage circuit that sequentially stores image signals of blocks composed of nxn pixels from a binarized image signal, and an AND circuit that assigns a predetermined order to each image signal and combines image signals between mutually adjacent orders, and an exclusive A tracking vector detection circuit that detects the order of signal level change points by a circuit including an OR circuit, and a block to be extracted next according to the tracking vector output from the tracking vector detection circuit by the circuit including an OR circuit. and an extraction direction indicating circuit for indicating the direction of the boundary line tracing circuit.