JPH02195485A - Image processor - Google Patents

Abstract

PURPOSE:To increase a processing speed by detecting an edge part for deciding a data value from binary data stored in an image memory, writing the address of the edge part in a memory and calculating the feature of a graphic included in an input image based on the edge address data. CONSTITUTION:An image signal inputted from a camera 1 is converted into a digital signal by an A/D converter 2, the digital signal is sent to the image memory 3 and the contents of the stored image signal is supplied to a binarizing circuit 4 at a prescribed timing. The circuit 4 discriminates the signal at a prescribed level, converts the signal into binary data '1' to '0' and sends the binary data to an edge measuring circuit 5, and a position to be changed from '1' to '0' or from '0' to '1' on a horizontal line is detected and added to an edge information memory 6. The stored data are sent to a CPU 8, a ROM 9, a RAM 10, etc., connected to a data/address bus 7 to calculate the area, centroid and inclination of a graphic. Thus, the hardware processing and software processing are combined to miniaturize the device.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of the Invention) This invention detects the edges of a figure included in an input image, and based on the address data, detects feature quantities such as the area, center of gravity, and slope of the figure. The present invention relates to an image processing device that can calculate .

(Summary of the Invention) The present invention provides an image processing device that detects the edge portion of a figure included in an input image and calculates feature quantities such as the area, center of gravity, and slope of the figure based on the address data. , edge detection and address data collection, which require a large amount of processing, are processed by a dedicated hardware circuit, and other calculations and other processing are performed by software processing. It is small in size and low in price.

(Prior art and its problems) Conventionally, in this type of image processing apparatus, there are two methods for extracting the features of a figure, which can be broadly classified as follows.

(1) After the image data is loaded into the memory, the feature amount of the figure is calculated by subjecting the data to software processing.

(2) Calculate the feature amount of the figure while importing image data from the outside or by processing these data with hardware after importing the data into memory.

However, in the case of software processing (1), the processing speed is slow, so there is a problem that it takes too much time to calculate.

In addition, in the case of (2) hardware processing, although the processing speed is faster, there are problems in that the device is large and expensive, and has poor versatility.

(Objective of the Invention) This invention was made to solve the above-mentioned problems, and its purpose is to calculate the feature amount of a figure in a short time, and at the same time to make the device itself small and inexpensive. The objective is to provide an image processing device that can.

(Structure and Effects of the Invention) In order to achieve the above object, the present invention provides an image processing device having a function of extracting features such as area double center of gravity and inclination of a figure included in an input image, which is stored in an image memory. an edge detection circuit having a hardware configuration, which detects an edge portion determined by a data value on a line forming an image from the binarized data, and writes the address of the edge portion into an edge information memory; The present invention is characterized by comprising an arithmetic processing unit that calculates the features of a figure included in an input image by software processing based on edge address data in an edge information memory.

According to this invention, detection of edge portions that require a large amount of processing and collection of their address data are processed by a dedicated hardware circuit, and other processing such as calculations is performed by software processing. , the processing speed can be increased, and at the same time, it can be configured to be small and low cost.

In addition, since the software of the calculation section can be changed,
This increases versatility in setting processing contents and increases the uses of the device.

(Example) Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the electrical configuration of an image processing apparatus to which the present invention is applied.

In the figure, an A/D converter 2 converts an image signal received from a camera 1 into a digital signal and sends it to an image memory 3.

The image memory 3 stores the input digital signal and sends the contents of the stored image signal to the binarization circuit 4 at a predetermined timing.

When the digitized image signal is manually input, the binarization circuit 4 discriminates it at a predetermined level, converts it into binary data of 1 or 0, and sends it to the edge measurement circuit 5.

The edge measuring circuit 5 detects the inverted part of data on one horizontal line that makes up the image, that is, the position where the data changes from 1 to 0 or from 0 to 1, and recognizes this as the edge part of the figure. , extracts the address (X coordinate) on the line corresponding to the edge portion and sends it to the edge information memory 6.

The edge information memory 6 stores input edge address data, and the stored data is sent onto the data/address bus 7 as appropriate.

On the other hand, CPU 8 . connected to data/address bus 7 .
ROM9. A microcomputer is configured by RAMl0 and the like. This microcomputer section reads the edge address data in the edge information memory 6 via the data/address bus 7, identifies the figure in the area designated as a window in the image, and calculates the area, center of gravity, inclination, etc. of the figure. The calculation process is performed.

In other words, in this device, a dedicated hardware circuit is responsible for detecting edges related to figures in an image and writing the edge address data, reads the edge address data, and based on that data, calculates the area, center of gravity, slope, etc. The process of calculating this is done by a microcomputer.

In this way, when detecting edges using a dedicated hardware circuit, unlike software processing which detects each pixel, it is possible to perform continuous processing, making it possible to detect edges with the greatest processing load. Processing speed can be significantly improved.

In addition, by using a microcomputer to handle the processing after writing the edge address data, the device does not need to be made larger, is less expensive, and processing settings can be easily changed by exchanging programs. .

Next, a procedure for calculating feature amounts of a figure using edge address data will be explained.

Generally, the area of a figure (AR
EA), center of gravity (XG, YG), principal axis angle (θ),
It is determined by the following formula.

AREA=ΣΣf (L, j)
(1) + ΣΣf (i, j) ・j Σ 1-1 = ΣΣf (1゜j) ・j2 (ΣΣf (i, j) ・i) 2 ” ΣΣ t, j ” O −ΣΣf (1゜j) ・j2 SUMX2−ΣΣf (i.

・12 (ΣΣf (i, j) ・j) 2 From now on, in order to make the equation easier to read, each term is redefined as follows.

At this time, equations (2), (3), (5), and (7) can be expressed as follows.

SUMX-ΣΣf (i.

j) f (i, j) in each of the above formulas is 0 or 1 because the figure is expressed by binary data.

Here, AREA is taken as an example of a specific binary image.

SUMX, SUMY, SUMXY, SUMX2゜SUM
Figure 2 shows how to determine Y2.

As shown in the figure, in the case of a binary image, if the number of calculated pixels in one line, the sum of weights in the X direction, and the sum of square weights in the X direction are found for each horizontal line, the Y coordinate can be calculated using a counter. By calculating as follows and adding them sequentially, AREA, SUM
X, SUMY, SUMXYSUMX2. Find SUMY2.

By substituting these values into equations (1) to (12), the area 1
The center of gravity and principal axis angle are determined.

Next, ■ the number of calculated pixels in one line, ■ the sum of weights in the X direction, ■
A method of calculating the sum of square weights in the X direction from edge address data will be explained.

When the edge coordinates of a pattern in one horizontal line are determined as shown in Figure 3, the sum of weights in the X direction (S n
) and the sum of square weights (Dn) are determined as follows.

5-(X2+X3+X4+X5) + (Xa +X9 +Xto) + (X12)D-
(X22+X32 +X, 2 +X52)+(Xa 2
+x92+Xto2)+ (X122) The sum and square weighted sum can be defined as follows.

In general, -OK-0, so if the edge coordinates (Xa and xb) of a pattern are determined, the load for that pattern is The information is input to the edge measuring circuit 5 for each horizontal line.
pixel) and stores its X coordinate in the edge information memory 6. The contents of this edge information memory 6 are stored by the CPU 10.
However, the order of reading at that time is from the right side.

Further, in this embodiment, in order to provide versatility, an arbitrary rectangular area on the screen is designated as a window, as shown in FIG. 4, and edge information inside the window is extracted. In the figure, only one line of the window is specifically shown.

This window is defined by the upper left coordinates (Xb, Yb) and size (Xs, Ys).

Here too AREA, SUMX, SUMY, SUMXY,
SUMX2. SUM Y2 is calculated as follows.

However, if En'=En+Xb (17) to (22) are substituted into equations (8) to (12), the area 9 center of gravity and principal axis angle can be determined.

Furthermore, for formulas (18), (20), and (21), if the calculation results are made into a table and stored in advance, the calculation time can be further reduced.

In this example, the edge address of a figure is determined by hardware, and the above calculation is performed by software from that information, so that the area 9 center of gravity,
Geometric features such as the principal axis angle can be determined.

Furthermore, by changing the software, other graphic feature quantities such as the peripheral length of the figure, the number of holes, and the number of pieces can also be determined at high speed.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the electrical configuration of an image processing device according to the present invention, FIG. 2 is an explanatory diagram of feature amount calculation of a figure, FIG. 3 is an explanatory diagram of a pattern on a horizontal line, and FIG. 4 is a drawing It is an explanatory diagram of a window set above. 1... Camera 2... A/D converter 3... Image memory 4... Binarization circuit 5... Edge measurement circuit 6... Edge information memory 7... Data/address bus 8...CPU 9...ROM 10...RAM Fig. 2 Fig. 1

Claims (1)

[Claims] 1. In an image processing device having a function of extracting features such as area center of gravity and slope of figures included in an input image, an image is extracted from binarized data stored in an image memory. A hard-configured edge detection circuit detects the edge portion for which the data value is determined on the constituent lines and writes the address of the edge portion into the edge information memory, and based on the edge address data in the edge information memory, An image processing device comprising: an arithmetic processing unit that calculates features of a figure included in an input image by software processing;