JPS59125474A - Picture processing method - Google Patents

Abstract

PURPOSE:To extract assuredly a pole and to recognize accurately an object by extracting a picture element having a larger difference in variable density level than any adjacent picture elements set horizontally, vertically or diagonally as a pole. CONSTITUTION:The variable density picture signal sent from a TV camera 1 is converted into a digital picture data signal by an A/D converter 2 and then applied with the primary or secondary differentiation through a space filter 3. The output of the filter 3 is supplied to a data shift circuit 4, and the circuit 4 stores and shifts the data signal needed for extraction of a pole. Then a data processing circuit 5 extracts a picture element having a larger difference in variable density level than any adjacent two picture elements which set horizontally, vertically or diagonally as a pole.

Description

DETAILED DESCRIPTION OF THE INVENTION (1) Technical Field of the Invention The present invention relates to an image processing method, and particularly to an improvement in an image processing method that is effective in extracting extreme points of a grayscale image.

(2) Background of the technology In general, in images obtained from TV cameras, etc., parts where the brightness changes suddenly are considered to be important points corresponding to line points on the object's surface. Connect the dots in a series of 9
This is an important process for pattern recognition of objects.

(3) Prior art and problems Based on the above-mentioned concept, the applicant has previously proposed the following image processing method (Japanese Patent Application No. 57-10471).
No. 4). For example, after subdividing the multilevel grayscale image obtained through preprocessing into pixel units, and comparing the grayscale levels of the pixel of interest and pixels adjacent to it in the horizontal and vertical directions, either When a pixel of interest out of two pixels lined up in the direction of is at a high level, the pixel of interest is extracted as an extremum. According to this method,
By finding the above-mentioned extreme points as a series of connected lines, it becomes possible to recognize the outline of the object.

However, in such an image processing method, for example, when recognizing the outline of overlapping objects or the outline of an object with corners, the pixels corresponding to the branching points of the outline are Since the pixel is located on two contour lines, if the respective contour lines were to extend horizontally and vertically in the vicinity of the pixel portion, the pixel corresponding to the branching point of the contour could not be extracted as an extremity. As a result, the outline of the object is interrupted, and the recognition of the object becomes inaccurate accordingly.

(4) Purpose of the Invention The present invention has been made from the above point of view, and its purpose is to eliminate discontinuities in the contours of objects;
Therefore, it is an object of the present invention to provide an image processing method that enables accurate object recognition.

(5) Structure of the Invention The basic structure of the present invention is to divide an image obtained from a receiver or the like into a plurality of pixels arranged in a matrix, and among the plurality of pixels, adjacent horizontal and vertical Furthermore, there is an image processing method in which a pixel having a larger difference in gray level than any two pixels lined up in the right diagonal direction is extracted as an extreme point.

(6) Embodiments of the Invention Hereinafter, the present invention will be described in detail based on embodiments shown in the accompanying drawings.

First, the basic principle of the extreme point extraction process according to the present invention will be explained based on FIG. Now, in order to simplify the explanation, when the density distribution in the one-dimensional direction of the original image is P as shown in FIG. First, as shown in FIG. 1(b), move the original image P by one pixel △ to the left.
The image pattern shifted by
As shown in , this left-shifted image pattern PL and the original image P are compared to extract a portion QL in which the density is greater than or equal to that of the original image P (hereinafter referred to as P≧Px), and the first As shown in figure (d), the original image P is moved one pixel △ to the right.
The image pattern shifted by the amount is set as PR, and Fig. 1(e)
As shown in FIG. 1, this right-shifted image pattern PR and the original image P are compared to extract a portion QR where the density of the original image P is greater than or equal to that of the original image P (hereinafter, this will be referred to as P≧PR). Then,
Since the sought extreme point R is considered to be at least as large as or equal to the pixels on both sides, it can be seen in Figure 1(f).
As shown in FIG.
can be obtained. However, P=PL=PR
In this case, the pole R does not exist and is therefore excluded.

Now, if we subdivide a two-dimensional image into a matrix, that is, into pixel units aligned in the horizontal and vertical directions, we can use the basic principle described above to divide the pixel of interest in the two-dimensional image into horizontal, vertical, and left-aligned pixels. It is possible to individually judge whether the point is the extreme point in the corner direction (multiple directions in the upper left with respect to the horizontal direction) or in the right diagonal direction (multiple directions in the upper right with respect to the horizontal direction). In this case, in order to determine whether or not the pixel of interest E is a pole in each direction, as shown in FIG.
C, D, F, G.

Window W is set with H and I.

FIG. 3 shows a specific algorithm for extracting extreme points using the window W described above.

First, to explain the case of determining whether or not a pixel of interest E in a two-dimensional image falls at a pole in the horizontal direction, the gray level of the pixel of interest E and the pixels F and F on both sides of this pixel of interest E are compared. , if E≧D (the density of pixel E is greater than or equal to the density of pixel A, the same applies hereafter), then 5.
L-1, otherwise 5L=O, and if E≧F, 5R=1, otherwise 5R=0, but E
If =D=F, the density level does not change, so 5N-
0, otherwise 5N=1. Then, an index S indicating whether or not the pixel of interest E is a pole in the horizontal direction
is the AND operation of SL, SR and SN, i.e. S = S
L X SRX SN is obtained, and when S=1, it indicates that the pixel E is a pole, and when S=0, it indicates that the pixel E is not a pole.

Next, to explain the case of determining whether or not the pixel of interest E of a two-dimensional image is a pole in the vertical direction, the pixel of interest E, the pixels B above and below this pixel of interest E, and the pixels B above and below this pixel E and Compare the gray level with H, and if E≧B then TL=1, otherwise TL=0, if E≧H then TR-1, otherwise TR=0, however,
If E=B=H then TN=0, otherwise TN=
Set to 1. Then, an index T indicating whether or not the pixel of interest E is an extreme point in the vertical direction is obtained by a logical product operation of TL, TR, and TN, that is, T:TL XTRXTN, and T
=1 indicates that the pixel E is the extreme point, and T-0 indicates that the pixel E is not the extreme point.

In addition, we will explain the case of determining whether or not the two-dimensional pixel E is a pole in the right diagonal direction (the same applies to the left diagonal direction, so it will be indicated in parentheses below). , substantially in the same way as described above, compare the gray levels of the pixel of interest E and the pixels C and G (A and I) adjacent to the pixel of interest E in the right diagonal direction (left diagonal direction), and find that E≧ If C(IN≧A), UL=1(Vr,=1),
Otherwise, Ur, = O (VL = 0), or if E≧G (E≧1), then UL = 1 (Vt, = 1
), otherwise UL=O (VL=0), but if E=C=G (E=A=I) then UN=12
(VN=ρ), otherwise UN=l (VN-/)
shall be.

Then, the index U(V) indicating whether or not the pixel of interest E is the extreme point in the right diagonal direction (left diagonal direction) is UL, UR
and UN (VL, VR and VN), i.e. U=ULXURXUN (V=Vt, XVRXVN
), and when U(V)=1 or O, it indicates that the pixel E is a pole or not a pole.

Furthermore, an index X indicating whether the pixel of interest E is a pole in any direction on the screen is obtained by a logical sum operation of SIT, U, and v, that is, X=S+T+U+V,
When X = 1, it indicates that pixel E is a pole in either direction, and when X-0, it indicates that pixel E is not a pole in either direction, and the value when b, x = i is set to a high level, and the value at X-0 is set to a low level to extract the extreme point. When the window W is moved on the screen using such an algorithm, all the extreme points on the two-dimensional image will be extracted.

An example in which the above-described image processing method is applied to an object recognition device will be shown below.

The basic configuration of this object recognition apparatus is shown in FIG. A spatial filter 3 that performs differential calculations, and data that stores and stores data signals necessary for pole point extraction processing among the digital image data signals from the spatial filter 3, and that shifts the data signals in synchronization with the horizontal scanning signal of the TV right camera. It consists of a shift circuit 4 and a data processing circuit 5 that performs calculations for pole point extraction processing based on data signals stored in the data shift circuit 4.

Now, as shown in Fig. 5, if a circular object 7a and a triangular object 7b are placed on the table 6, overlapping each other, and these objects 7a and 7b are photographed with the TV right camera, this A grayscale image signal is output from the TV camera 71 in synchronization with the horizontal scanning signal of the TV right camera, and the multilevel grayscale image is subdivided into pixel units. The output signal from AD converter 2
The image data is converted into, for example, an 8-pin digital image data signal. In this case, as shown in Figure 6,
If we pay particular attention to the branching part of the outline of objects γa and 7b (the part surrounded by the dashed line) in the image taken by the TV right camera, the digital grayscale image based on the digital image data signal appears as shown in FIG. , each data value indicates the gray level at each pixel. Then, the output signal from the AD converter 2 passes through a spatial filter 3 and is converted into a digital image data signal with emphasized gray level. In this case, the digital gray image appears as shown in FIG. 8, for example. It should be noted that each data value in FIG. 8 indicates the gray level of each pixel.

As shown in FIGS. 4 and 9, the data shift circuit 4 stores data corresponding to 2N+3 pixels, assuming that the number of pixels arranged in the horizontal direction of the image screen shot by the TV left camera is N. For example, registers R6 to R6c+ each have an 8-bit data width, and shift registers S1 . It is composed of S2,
The data stored in the registers R61 to Re9 correspond to the pixels A to A shown in window W in FIG. That is, as shown in red in FIG. 10, horizontal scanning lines hl, h2, . When scanning is performed sequentially with h3..., the scanning by the scanning lines hx and h2 ends, and at the b floor where the scanning of the first two pixels by the scanning line h3 is completed, the data shift circuit 4 receives 2N- Three pieces of pixel data are stored in registers Rel to Re9, respectively.
33 (pixel located in the third row and third column on the photographed image screen,
The same shall apply hereinafter), g32. g31. g23...
The data corresponding to gll is stored and each data is used as data for pixels I to A of window W. When the scanning of the pixel g34 by the scanning line h3 is completed, the data corresponding to the pixel g34 is newly stored in the register R61, and at this stage, the data in each register is not entirely shifted. In addition, pixels g34 and 7 are stored in registers RBt to Re9, respectively.
g33, g32.

g24...The data corresponding to g12 is not stored and the window W is moved one column to the right. Further, as the scanning by the scanning line h3 progresses, the window W shifts to the right in synchronization with the scanning signal, and as the scanning by the scanning line h4 progresses, the window W shifts to the right in synchronization with the scanning signal.
moves down one line, then moves sequentially from left to right 1-
Teak.

In this way, the window W moves sequentially as the scanning line m scans, and at the end of the scanning line m, the window W moves across the entire range of the photographed image pixel by pixel. , this means that all the data necessary for the pole extraction process will be loaded.
Each data is sequentially sent to the data processing circuit 5.

As shown in FIGS. 4 and 11, the data processing circuit 5 also includes a horizontal comparison unit 5a that compares data between the pixel E of interest and the pixels F and F adjacent to the pixel E in the horizontal direction. , a vertical comparison unit 5b that compares data between the pixel of interest E and pixels B and H adjacent to the pixel of interest E in the vertical direction; and a pixel C of interest and the pixel C adjacent to the pixel of interest E in the right diagonal direction. and G, and a left diagonal comparison unit 5d that compares data between the target pixel E and the pixels A and G adjacent to the target pixel E in the left diagonal direction. and an output section 5e that outputs the comparison result, and a comparator C0
1.Co8, and gate AN1. AN2, orday) OR1...OR9 Nantes Gate NA1...
・Equipped with a peak register PM that outputs the data when the data of NA4 and the pixel E of interest is sent and the pixel of interest E is at the extreme point, and the circuit is configured to execute the algorithm shown in FIG. 3. . That is, comparator C(H
compares the data between pixels E and D, and when E≧D, [1
The comparator CO2 outputs pixel E (!:
Comparator Co3 compares data between pixels E and B and outputs "1" when E≧F, and comparator Co3 compares data between pixels E and B and outputs "1" when E≧B. The comparator CO4 compares the data of pixels E and H and outputs rlJ when E≧H, and the comparator C□s compares the data of pixels E and H.
The comparator co6 compares the data of pixels E and C and outputs "1" when E≧C, and the comparator co6 compares the data of pixels E and G and outputs "1" when E≧G. The comparator CO7 compares the data of pixels E and A and outputs rlJ when E≧A, and the comparator c08 compares the data of pixels E and ■. When E≧1, rlJ is output. If the comparators COI and CO2 both output "1" and E = D = F, in other words, if pixel E is the horizontal pole, the AND gate ANI outputs rlJ, and the others do the same. If pixel E is the vertical pole, the AND gate AN2 is "
1", and if pixel E is the extreme point in the right diagonal direction, AN3 outputs "1", and if pixel E is the extreme point in the left diagonal direction, AND gate AN4 teeth"
1" is output. Therefore, OR9 should output "1" when pixel E is a pole in either the horizontal direction, vertical direction, or left/right diagonal direction, and the pole point can be extracted using this output signal as index X. It will be done. At the same time, when OR9 outputs IJ, the data at the extreme point will be obtained as is from the peak register PM.

In this case, since the digital grayscale image data is shinted in the object recognition device in this example, it is possible to extract the extreme points in real time.

By pressing in this way, the extreme points of the grayscale image are extracted, but the 8th
When the poles are actually extracted from the digital grayscale image shown in the figure, the points surrounded by circles and squares in the figure are the poles. In this case, the poles surrounded by circles are extracted as poles in the horizontal or vertical direction, whereas the poles surrounded by squares are not extracted as poles in the horizontal or vertical direction, but in the left or right diagonal direction. This is the first time that it has been extracted as an extreme point. The pole part surrounded by the square mark above is, for example, the overlapping object γa, 7
By extracting the extrema of this part, the discontinuity (d
) (see FIG. 12) can be reliably eliminated, and the contours of overlapping objects can be recognized accordingly.

In the above embodiment, the present invention is applied to the digital grayscale image that has passed through the spatial filter 3r, but the present invention may also be applied to the digital grayscale image directly from the TV left camera. Further, the specific means using the image processing method according to the present invention is not limited to the above-mentioned embodiment, and it goes without saying that the design may be changed as appropriate.

(7) As described above in detail, according to the image processing method of the present invention, an image obtained from a receiver or the like is divided into a plurality of pixels arranged in a matrix, and the plurality of pixels are Among the pixels, a pixel with a larger difference in gray level than any two neighboring pixels arranged horizontally, vertically, and left/right diagonally is extracted as a pole, so that the pole can be extracted reliably. Even when recognizing the outlines of overlapping objects, etc., there is no risk that the outlines of the objects will be interrupted, and the objects can be recognized more accurately.

[Brief explanation of the drawing]

FIGS. 1(a) to (f) are explanatory diagrams showing the basic principle of extracting the extreme points of a -dimensional image; FIG. 2 is an explanatory diagram showing an example of a window used when extracting the extreme points of a two-dimensional image;
Fig. 3 is an explanatory diagram showing the algorithm of the image processing method according to the present invention, Fig. 4 is a schematic diagram showing an example of an object recognition device to which the image processing method according to the present invention is applied, and Fig. 5 is a photograph taken by the TV right camera. Figure 6 is a diagram showing an image taken by the TV right camera, Figure 7 is a diagram showing an example of digital image data obtained by digitizing the image taken by the TV right camera, and Figure 8 is a diagram showing a spatial filter. FIG. 9 is a circuit diagram showing a specific example of the data shift circuit shown in FIG. 4, and FIG. 10 is a diagram showing an example of the digital image data that has passed through the
Explanatory diagram illustrating an example of the camera photographed image screen review method, Part 1
1 is a circuit diagram showing a specific example of the data processing circuit shown in FIG. W...Window A~■...Pixel 1...
TV right camera 2...AD converter 4...Data shift circuit 5...Data processing circuit [@ (0) Year/month diagram Figure 1 (bl FC) Figure 1 (f LIIOR Figure 2 Figure 1 Figure s Figure 9 Figure toyA

Claims (1)

[Claims] An image obtained by a television receiver or the like is divided into a plurality of pixels arranged in a matrix, and from among the plurality of pixels, any two of the adjacent pixels arranged in the horizontal, vertical, and left/right diagonal directions are An image processing method characterized by extracting pixels having a large difference in gray level as extreme points.