JPH087113A - Image feature extraction method - Google Patents

Abstract

(57) [Summary] [Purpose] To enable implementation with a simple circuit configuration. [Structure] F / F2 outputs negative phase output "1" in the initial state.
Therefore, when the binarization processor 33 first sets the output to “1”, the gate 3 outputs a latch command for the F / F 4, and the Y coordinate data at that time is fetched in the F / F 4. Since F / F2 sets the reverse phase output to "0",
/ F4 holds and outputs the minimum value on the Y coordinate. Meanwhile F / F
In 5, every time the binarization processor 33 sets the output to “1”, Y
Although the coordinate data is fetched, the binarization processor 33 sets the output to "1" at the end just before the vertical scanning passes the target point, so that the F / F5 has the maximum on the Y coordinate. The value is retained.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for extracting a feature amount of an image, in which the contour on the Y coordinate of one article is extracted from the image data of one article.

[0002]

2. Description of the Related Art In a raster scan system, for example, as shown in FIG. 2, an upper left corner of a rectangular screen is used as a scanning starting point, and an image of one screen is obtained by horizontal scanning from left to right and vertical scanning from top to bottom. Although this is a method of acquiring a signal, the position within the screen is indicated by XY coordinates with the horizontal scanning direction as the X axis and the vertical scanning direction as the Y axis.

Here, in the image feature amount extraction method of the present invention, as shown in FIG. 2, a target point 1 which is an article is displayed in the screen acquired by such a raster scan method.
This is a method of extracting (detecting) the contour (minimum value, maximum value) on the Y coordinate of the target point 1 when only the target point 1 exists. As a feature amount extraction method of this type of image, a method shown in FIG. 3 is conventionally known.

In FIG. 3, the image signal picked up by the video camera 31 includes only one article. The image signal acquired by the video camera 31 is digitized by the A / D converter 32.

In the binarization processor 33, the A / D converter 32 is used.
Of the output value and the threshold value are compared, and if the output value exceeds the threshold value, "1" is output, and if it does not exceed the threshold value, "0" is output.
Is output. The output of the binarization processor 33 is given to the control terminals G of the two comparators (34, 35).

In the comparator 34 (35), the Q output of the flip-flop (F / F) 36 (37) is applied to one input terminal A, and the output of the A / D converter 32 ( Y
Coordinate data) is applied, and the magnitude relationship between both inputs is compared every time "1" is input from the binarization processor 33. When the magnitude relationship between both inputs is A> B (A <B), the F / F 36 The signal of the latch command is output to the latch input terminal of (37). This 34
When the output of the binarization processor 33 becomes "1", the two comparators 35 and 35 simultaneously start the comparison operation of both inputs.

The F / F 36 (37) has an A / D at the input terminal D.
The output (Y coordinate data) of the converter 32 is applied, the Y coordinate data is fetched in response to the latch command, and the Y coordinate data is output from the output terminal Q. This F / F36 and 37 are
Both of them exist for the number of bits of the Y coordinate data, but in the initial state, the F / F 36 sets all the Q outputs to "1" to prepare for the input of the latch command signal, and the F / F 37 sets all the Q outputs to "0". To prepare for inputting the latch command signal.

In the above configuration, when the binarization processor 33 first outputs "1", the condition of A> B is satisfied in the comparator 34, the F / F 36 is activated, and Y is output to its output terminal.
Coordinate data is output. Then, for the subsequent input of "1", if the condition of A> B is satisfied in the comparator 34, F
The Y coordinate data input this time is loaded into / F36. In this way, the minimum value on the Y coordinate is detected and output to the Q output of the F / F 36. In many cases, the Y coordinate data captured when the binarization processor 33 first outputs "1" is the minimum value.

On the other hand, in the comparator 35, since the initial value of the Q output of the F / F 37 is "0", the condition of A <B is satisfied until the vertical scanning passes the target point 1, and until then, 2 The Q output of the F / F 37 is updated every time the output of the digitization processor 33 becomes "1", and the updating operation is stopped after A> B. That is, to the Q output of the F / F 37, the Y coordinate data immediately before the vertical scanning passes the target point 1 is output as the maximum value on the Y coordinate.

[0010]

In the conventional image feature amount extraction method, the minimum value and the maximum value are detected by comparing the previous value and the current value, so that a comparator is required, but many comparators are required. Not only does the circuit scale increase because it is a functional circuit, but there is also the problem that the number of patterns becomes enormous because the wiring pattern for the number of bits is also required for the comparator.

The present invention has been made in view of such conventional problems, and an object thereof is to provide an image feature amount extraction method capable of reducing the circuit scale and the number of wiring patterns. .

[0012]

In order to achieve the above-mentioned object, the image feature amount extraction method of the present invention has the following configuration. That is, the image feature amount extraction method of the present invention digitizes an image signal obtained by picking up one article by the raster scan method, and digitizes the image signal from the digital signal when the binarized data is "1". Image feature amount extraction method for extracting the contours (minimum value, maximum value) of the article in the vertical scanning direction; the digital signal when the binarized data first becomes "1" is the minimum value; The maximum value is the digital signal at the last "1" when the digitized data continues to be "1".

[0013]

Next, the operation of the image feature amount extraction method of the present invention configured as described above will be described. In the present invention, the digital signal when the binarized data first becomes "1" is detected as the minimum value, and the digital signal at the last "1" when the binarized data continues "1". The signal is detected as the maximum value.

Therefore, the circuit for implementing the method of the present invention can be constituted by the flip-flop and the gate, and the circuit scale can be reduced. Further, the wiring pattern can be reduced because there is no comparator.

[0015]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 shows an example of a circuit for implementing the image feature amount extraction method of the present invention. In FIG. 1, the implementation circuit of the method of the present invention has an F / F2, a gate 3, an F / F4, and an F / F5 provided in a stage subsequent to the binarization processor 33.

The output of the binarization processor 33 is supplied to the input terminal J of the F / F2, one input terminal of the gate 3 and the latch input terminal of the F / F5. The Q-bar output of the F / F2 (hereinafter referred to as "negative phase output") is supplied to the other input end of the gate 3, and the output of the gate 3 is supplied to the latch input / output end of the F / F4. The output (Y coordinate data) of the A / D converter 32 is supplied to the input terminal D of the F / F 4 and the same 5.

In the above configuration, the F / F2 has the anti-phase output set to "1" in the initial state. Therefore, when the binarization processor 33 first sets the output to "1", the F / F2 outputs from the gate 3 to the F / F2.
The latch command of F4 is output, and the Y coordinate data at that time is fetched in F / F4. Then, the F / F2 sets the negative phase output to "0" by the first input of "1" from the binarization processor 33, so that the binarization processor 33 thereafter outputs "1". Also, the gate 3 closes the gate, the latch command is not output to the F / F 4, and the F / F 4 holds and outputs the Y-coordinate data that is initially captured. As is clear from FIG. 2, the Y coordinate data that is first captured is the minimum value on the Y coordinate.

On the other hand, in the F / F5, the Y coordinate data is fetched each time the binarization processor 33 sets the output to "1".
The value processor 33 finally sets the output to “1” until just before the vertical scanning passes the target point 1. Therefore, after all, F
The maximum value on the Y coordinate is held in / F5.

Since this embodiment circuit is composed of a single-function flip-flop and a gate, it can be constructed very simply.
The number of wiring patterns is also reduced because there is no comparator.

[0020]

As described above, according to the image feature amount extraction method of the present invention, the digital signal when the binarized data first becomes "1" is detected as the minimum value, and the binarized data is detected. Since the digital signal at the last "1" when "1" continues to be detected as the maximum value,
A circuit for implementing the method of the present invention can be composed of a flip-flop and a gate, and the circuit scale can be reduced. Further, there is an effect that the wiring pattern can be reduced because there is no comparator.

[Brief description of drawings]

FIG. 1 is a circuit diagram showing an example of a circuit that implements an image feature amount extraction method of the present invention.

FIG. 2 is a diagram showing an example of a screen configuration by a raster scan method and an example of a target of feature amount extraction of the method of the present invention.

FIG. 3 is a circuit diagram showing an example of a circuit that implements a conventional image feature amount extraction method.

[Explanation of symbols]

 1 Target Point 2, 4, 5 Flip-Flop (F / F) 3 Gate 31 Video Camera 32 A / D Converter 33 Binarization Processor

Claims (1)

[Claims] 1. An image signal obtained by picking up an image of one article by a raster scan method is digitized, and the contour of the one article in the vertical scanning direction is obtained from the digital signal when the binary data of the image signal is "1". In the image feature amount extraction method for extracting (minimum value, maximum value); the digital signal when the binarized data first becomes "1" is set as the minimum value; the binarized data continues to be "1" The maximum value is the digital signal at the last "1" when performing the following;