JPS63180178A - Edge emphasizing device - Google Patents

Abstract

PURPOSE:To realize the uniform edge emphasis of an entire area with elimination of background noise by obtaining a projection distribution from the title histograms of each medium area and deciding an executing area from said projection distribution to perform prescribed emphasis processing within said executing area. CONSTITUTION:The pictures of a TV camera 2 are digitized and inputted to a picture memory 4. A differential arithmetic part 5 calculates the differential pictures of the original ones and inputs them to a differential picture memory 6. An edge tilt angle arithmetic part 7 and an an edge intensity arithmetic part 8 calculates an edge tilt angle theta and edge intensity (w) from the differential pictures and store them in an edge tilt angle memory 9 and an edge intensity memory 10 respectively. An edge emphasis arithmetic circuit 11 calculates the edge tilt angle histogram for each minor area to obtain the tilt angle histogram for each medium area of division A and B respectively. Then the peak value thetap of the histogram is detected for each medium area and the projection distribution of the intensity (w) is obtained within the medium area rotated in response to the value thetap for the edge within the half value width of the value thetap. Then an emphasis area corresponding to the value thetap is decided and the edges are emphasized within the emphasis area together with other edge intensities set at zero respectively.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a device that performs edge enhancement processing to extract a contour from a grayscale image of an object. Note that such a device is suitable for use in a robot vision system that recognizes the shape of an object from its contour and measures its position.

[Conventional technology]

Conventionally, various differential operators (5obel, Roberts
and Roblnson's operators) are known.

[Problem that the invention seeks to solve]

However, even with this method, there is a problem that edges with small differential values and noise due to background texture remain. Further, noise removal and edge extraction are performed in contour extraction processing and recognition processing, but these are not always sufficient. There are problems such as blanks, residual noise line segments, and parts of contour lines that cannot be extracted.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide an edge enhancement device that can remove background noise, correctly extract edges with small differential values if they are part of a contour, and enhance the edges.

[Means for solving problems]

Differentiate the grayscale image obtained by imaging the object and quantizing it,
There are two ways to divide the effective screen: a calculation means that calculates the edge tilt angle and its intensity; The peak value is detected from the tilt angle histogram of each middle region, and the peak value is detected from the tilt angle histogram of each middle region. a determining means for determining the range of the corners; a calculation means for rotating the middle area in a direction corresponding to the tilt angle peak value and calculating a projection distribution in a direction perpendicular to the edges; and executing edge enhancement processing from the projection distribution of the edges. A determining means for determining the area to be emphasized and a processing means for emphasizing the edges near the emphasis area and setting the intensity of the other edges to zero are provided.

[Effect]

A grayscale image of the object is input and differential calculation is performed to calculate the edge inclination angle and edge strength. After that, change the screen to 2NX
It is divided into 2N small areas and divided into A division and B division.
A type of medium area (2×2 small area) division is performed, and an angle histogram is calculated for each division. Edges with the same inclination angle that exist within each middle region are detected from the peak value of the angle histogram. Furthermore, by rotating the middle region by the extracted tilt angle and calculating the projection distribution in the direction perpendicular to the edge direction within that region, the edges of the object and noise edges are separated and only the edges of the object are emphasized. Process. Also, by performing this emphasis processing on the A and B divisions,
Even if an edge on the periphery of the middle region is determined to be noise and is not emphasized in one division, K is set so that it can be emphasized in the other division.

If I were to break it down into bullet points, it would be as follows.

(1) Divide the effective screen into 2N x 2N small areas and
A histogram of edge inclination angles within the ×2 small region and the tb middle region is calculated to detect representative edge inclination angles existing within the middle region.

(2) Rotate the middle region in a direction corresponding to a typical tilt angle within the middle region, calculate its projection distribution in the direction perpendicular to the edges, and calculate the area where edges with the same tilt angle exist (highlighted region ) to determine.

(6) Enhancement processing is performed on the edges within the edge enhancement area obtained for the representative edges of each medium area.

(4) By performing the same processing not only for the A division but also for the B division, the emphasis processing is performed uniformly on all edges.

〔Example〕

FIG. 1 is a block diagram showing an embodiment of the present invention.

In the figure, 1 is an object, 2 is an imaging device such as a television camera, 3 is an analog/digital (A/D) converter,
4 is an image memory, 5 is a differential calculation section, 6 is a differential image memory, 7 is an edge slope calculation section, 8 is an edge strength calculation section, 9 is an edge slope memory, 10 is an edge strength memory, and 11 is an edge enhancement calculation section. Chiru.

An object 1 is imaged by a television camera 2, and the video signal thereof is stored in an image memory 4 as a digital value by an A/D converter 3. The image memory 4 stores the density value f(x, y
) is stored, for example, as 8-bit data. The coordinates (x, y) of the image are expressed using the X, Y coordinate system shown in FIG. The density value f(xty) of the pixel of interest (x, y) is
0≦f(x,y)≦255. This image person is differentiated by a differential calculation unit 5, and the result is stored in an image memory 6.

There are two directions for differentiation: the X direction and the Y direction, and their negative order differentials are expressed as Δxf and Δaf, respectively. Here, a 5obel differential operator as shown in the figure is used. The X-direction differential D1 shown in the figure (a) and the y-direction differential D2 shown in the figure (b) are expressed by the formula Δzf-(f(x+Ly-1)+2・f(x+l,y)
+f(x+1.y+1)) (f(x-1,y-1)
+2-f(X-1,y)+f(X-1ty+1))...
...(1a) Δyf-(f(x+1.y+1)+2・f(x,y+1
)+f(x-1,y+1))(f(x+1.y-1)
)+2・f(x,)'-1)+f(x-1,)'-1)
)...(1b) becomes.

The differential values (Δxf, jyf) are illustrated in FIG. 4. As is clear from the figure, the pixel of interest P e
The inclination angle θ and the strength W of the edge vector (x*y) are expressed by the following equation.

”= VE8 to F〒7.(,f)” −・4(3
) Here, the unit vector (
u, v), find θ using a table of inclination angles. If we want to find l with 1 degree precision, then (u,
It is sufficient to have t-560 pairs of values of y). The unit vector is as follows.

The edge tilt angle calculation unit 7 calculates the tilt angle θ from the unit vector (u, v) using an angle table, and stores the result in the edge tilt angle memory 9. The edge strength calculation unit 3 is (
3) Calculate the edge strength [w according to the equation and store it in the edge strength memory 10. The edge enhancement calculation unit 11 accesses the edge inclination angle memory 9 and the intensity memory 10 to perform edge enhancement processing.

Now, the effective screen of object 1 is 512 (horizontal) x 512 (
Vertical) 1m element, as shown in Figure 5, 2Nx
Divide into 2N small areas RSK. Each small area is shown in FIG. 6(a) and will be referred to as A division. The area divided by the broken line is a small area R5. In a small area of 2Nx2N,
FIG. 6(b) shows the (2N-2)×(2N-2) area, excluding the outer small area, divided into middle areas. The division shown in FIG. 6(b) is referred to as B division. next,
A histogram of the edge inclination angle θ is obtained for each small region. Further, when the θ histogram Hθ for each middle region of the A division and the B division is calculated, as shown in FIG. 7, there are peak values (Pθ1, Pθ2, etc.) corresponding to the contour. Then, the peak value of the tilt angle histogram is determined for each middle region. Only one peak is shown as KPθ in FIG. Next, the slope angle θ corresponding to the peak value Pθ of the θ histogram is
Pf: Find and further correspond to half value Pθ/2 (θPL+
θPU) to find the half width Wθ. Intensity histogram Hwt− for the same middle region as the tilt angle histogram
The result is as shown in Figure 9. Since the W histogram in the middle region of and does not have information on the location of edges, it cannot be distinguished from noise. Therefore, as shown in FIG. 10, the middle region Rc of interest is rotated by the inclination angle θP of the detected edge, and the projection distribution Sw of the edge is obtained in the edge direction Ce and the vertical direction Y'. From the projection distribution of edges in the θP direction, the region Re where the edges of the contour exist can be determined.

The projection distribution Sw of θP is shown in FIG. The position Xp is found for the peak value PxK of the projection distribution, and the emphasis area (Xpt, m
) can be determined. The inclination angle θ of the edge is [θPI, tθP
6 in the range of the UI, and only for edges in the emphasized region Re in the rotated middle region R1, the edge strength w
twsO value, and for other edges, the edge strength W is set to 10 m. This completes the edge enhancement of the contour in a particular direction.

The above emphasis processing is shown as a flowchart.
It will look like the figure.

(1) Input the image from the TV camera into the image memory 4 (
■Reference).

(2) Calculate the differential image of the original image (by differential calculation),
Input to the differential image memory 6 (see ■).

(3) Calculate the edge inclination angle θ and the intensity W from the differential image 6 and store them in the edge inclination angle memorator and the edge intensity memory 10 (see ■ and ■).

(7) Middle region rotated corresponding to each peak value θP (
9) By processing the peak value θpK of the edge inclination angle more than required, noise edges can be removed and the entire surface can be uniformly emphasized.

〔Effect of the invention〕

According to this invention, the effective screen is divided into 2N×2N small regions, a histogram of the edge inclination angle θ is calculated for each 2×2 small region, and every medium region. Since the direction is detected, even weak edges can be detected. Furthermore, from the detected edge inclination angle θP, the projection distribution of edges near θP in the rotated middle region is calculated to determine the area where edges exist, so edges with noise are removed from edges in the same direction. can do. Furthermore, since the above steps are performed for the middle regions of both the A and B divisions, edges can be uniformly enhanced in the entire region.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is an explanatory diagram for explaining the relationship between the object image in the imaging screen and differential processing, and FIG. 3 is a block diagram showing a differential operator. Fourth
The figure is an explanatory diagram for explaining edge strength and its slope.
Fig. 5 is an explanatory diagram for explaining an example of a method for dividing an effective screen into small areas, Fig. 6 is an explanatory diagram for explaining two types of medium area dividing methods, and Fig. 7 is an explanatory diagram for explaining an example of a method for dividing an effective screen into a small area. A graph showing the angle histogram, Fig. 8 is an explanatory drawing for explaining the peak value and half-width of the tilt angle histogram, Fig. 9 is a graph showing the intensity histogram, and Fig. 10 is an explanatory drawing for explaining the intensity distribution within the head region. FIG. 11 is an explanatory diagram for explaining the intensity projection distribution, and FIG. 12 is a flowchart showing the edge enhancement processing flow. Code explanation 1...Target, 2...TV camera,
3... A/D converter, 4... Image memory, 5... Differential operation section, 6... Differential image memory, 7... ...Edge slope calculation section, 8...
...Edge strength calculation unit, 9...Edge inclination memory, 10...Edge strength menu, 11...
...Edge emphasis calculation section. Agent Patent attorney Akio Namiki Agent Patent attorney Kiyoshi Matsuzaki 111 Figure g2 Figure Atsushi 3 Figure (a) (b)
Figure 4 Figure 5 Figure S @ Figure 6 (a) et al.

Claims (1)

[Claims] Differentiating a grayscale image obtained by imaging and quantizing an object,
A calculation means for calculating the edge tilt angle and its strength; a calculation means for dividing the effective screen into a plurality of small regions and calculating a histogram of the tilt angle for each small region;
a calculating means for calculating a tilt angle histogram for each middle region; , a determining means for determining the range of the tilt angle; a calculation means for rotating the middle area in a direction corresponding to the peak value of the tilt angle and calculating a projection distribution in a direction orthogonal to the edge; and edge enhancement processing from the projection distribution of the edge. 1. An edge enhancement device comprising: determining means for determining an area in which to perform the process; and processing means for emphasizing edges near the emphasis area and reducing edge strength to zero for other edges.