JPH01276278A - Contour line persuit processing system for multilevel image - Google Patents

Abstract

PURPOSE:To efficiently execute a contour line persuit processing to a multilevel image in a short period at a high speed by investigating the position information of all feature points and contour points with one scanning to the multilevel image made into an object and executing the contour line persuit processing based on the position information. CONSTITUTION:Image data made into the object are scanned by using the mask of 2X2, the contour point and feature point are obtained, and 1 and 2 are written onto a map in case of the contour point and feature point, respectively. The scanning is executed from the head of the map, and when a feature point K is found, a contour persuit is started from the position of the image data corresponding to the feature point K. A contour point on the map, in which the persuit is completed, is deleted, the scanning is executed again on the map from the position of the feature point K, and the action is repeatedly executed until it is made impossible to find the new feature point. Thus, a scanning frequency to an object image is reduced, and a processing time is remarkably shortened.

Description

[Detailed Description of the Invention] (Technical Field) The present invention relates to contour tracing processing for extracting all the components of closed curves created by figure boundaries, and in particular, the present invention is concerned with contour tracing processing for extracting all components of closed curves created by figure boundaries, and in particular, the present invention is designed to perform contour tracing processing at high speed for multivalued images. This paper relates to a contour tracing processing method for multivalued images.

[Prior art]

An example of a conventional contour tracing processing method for binary images will be explained using the figure shown in FIG.

In this process, we first start by finding the tracing start point for contour tracing, and in the process of tracing in sequence, we mark the traced points one after another as we proceed with the tracing. After completing one round of tracking, the first
Find one contour line as shown in Figure 0. By repeating this procedure, all contour lines in the target image are extracted.

Below, the algorithm for tracing the contour line in the case of 8-connection will be explained step by step with reference to FIG.

■ First, the pixel at the top left of the screen is set as the initial point of rask scanning to search for the tracking start point.

■ Continuing mask scanning, find 11 pixels that do not have a tracked mark yet, and select these 11 pixels (boundary points).
A tracked mark is attached to the tracked point, and this is set as the tracking start point X0.

In addition, at this time, if the above 1 = pixel is not in the image,
Terminate the algorithm.

■ Tracking starting point X above. The 8th neighborhood seen in the center is the 11th
In figure (a), starting from the element with number 1 and moving counterclockwise (
When tracing the inner contour, move clockwise from element number 4)
The first 11 pixels encountered (the element numbered 1 in the figure) are set as Xl, and this is set as the next boundary point.

At this time, if no pixel of 11 is encountered in numbers 1 to 8, this point is regarded as a meaningless noise element (isolated point) and the process returns to step (2) again.

■ Mark the above boundary point ) Numbers 1 to 7 (at this time, the previously determined boundary points are ○
The first 11 pixels encountered are set as X2, and this is the next boundary point (in the figure, the position is)
shall be.

■ Repeat the procedure (■) for finding X2 from X[ above, and find adjacent boundary points one after another.

At this time, if x, , , = x, , x, 1 = x, then X, , X, . . .

■ Next, return to ■ to find another contour line.

Note that when restoring the original figure from the contour line obtained as described above, basically the elements between the outer contour line and the inner contour line may be integrated into 11 pixels. Moreover, although the case of 8-connected contour lines has been described above, it may be performed using 4-connected contour lines.

However, when the above contour tracing processing method is applied to a multivalued image, it is divided into many segments because tracing does not start from a feature point, and each time a line segment or closed curve is traced, the beginning of the target image is The problem is that the number of tracking increases and a lot of processing time is required because the tracking has to be scanned to find the starting point (feature point).

〔the purpose〕

The present invention has been made in order to solve the above-mentioned conventional problems, and an object of the present invention is to provide a contour line tracing processing method for multivalued images, which performs contour line tracing processing in a short time and at high speed. .

〔composition〕

In order to achieve the above object, the present invention detects all feature points (
The system is characterized by examining the positional information of the intersection points (intersections and endpoints) and contour points, and tracing the contour line based on the positional information, thereby reducing the number of scans for the target image and significantly shortening the processing time. be.

Hereinafter, a detailed explanation will be given based on one embodiment of the present invention.

First, an algorithm for contour tracing processing for a multivalued image according to the present invention will be explained step by step.

(2) Obtain contour information and all feature points Ki (i=1 to 2) from the target image data using a 2×2 mask, and create a map (MAP) of the contour information and feature points.

■ If there are no feature points, move to ■.

■ Contour tracing is performed on the image data using a 2×2 mask from the position on the image data corresponding to the intersection point and end point Ki on the map, and the corresponding contour information on the map is sequentially erased. Continue until you reach the feature point.

(2) When there is no more contour information starting from the feature point Ki on the map, scan the map from the position of the feature point Ki to find the next feature point Ki+1.

■ Return to ■.

■ Search for new contour information by starting scanning from the position of the previous contour information on the map.

However, the first scan should be at the beginning of the map. Then, if there is no contour information, the algorithm ends.

(2) Contour tracing is performed on the image data using a 2×2 mask from the position on the image data corresponding to the found contour information, and the corresponding contour information on the map is sequentially erased.

■ Return to ■.

In the above algorithm, (1) to (2) are tracking processes for contour information that has feature points, and (2) to (2) are tracking processes for contour information (closed loop information) that does not have any feature points.

Next, a specific method of algorithm (2) described above will be explained.

As shown in Figure 1, a 2x2 mask M is placed on the grid points of the image.
The target image data is scanned in the direction of (a) -> (b) -> (C) in the figure so that the center of the image is aligned with the center of the image.

In addition, in this case, the bottom part of the above image is shown in (d) to (
The above mask M is applied and scanned in the form shown in g).

Then, parts that satisfy any of the conditions (1) to (3) described below are extracted as feature points, and (
A portion that satisfies the condition 4) is extracted as a contour point, and 2 is written on the map if it is a feature point, and 1 is written if it is a contour point.

The conditions (1) to (4) above are shown below.

Condition (1): A part where the pixels in the mask M have three or more different values.
This is a case corresponding to b).

Condition (2): The part where the pixels in the mask M have two types of values and the two sets of diagonal pixel values are the same This condition is satisfied, for example, in the case shown in Figure 2 (C). .

Condition (3) The bottom part of the two images, that is, (e) ~ in Figure 1
In the form shown in FIG. 2(g), a portion of the mask M with different pixel values satisfies this condition, for example, in a case corresponding to FIG. 2(d).

Condition (4): Other than the above-mentioned conditions (1) to (3), parts of the mask M with two types of pixel values satisfy this condition, for example, in cases corresponding to (e) in Fig. 2). It is.

Note that the information in the map corresponds to the information on the grid points of the image data shown in FIG. 3(a), as shown in FIG. 3(b).

Based on the method described above, the map shown in FIG. 6(b) is created from the image data of FIG. 6(a).

In addition, ■ on the map in FIG. 2B indicates feature points, and ■ indicates contour points.

Next, a specific method of contour tracking in the above-mentioned algorithm (2) will be explained.

In this case, as shown in FIG. 7, the map shown in FIG. 7(b) is scanned sequentially from the beginning to find the feature point ■, and the image data shown in FIG. Tracking starts from the position of the feature point in .

Then, as shown in FIG. 4(a), priorities are assigned clockwise in the tracking direction from the feature point, and tracking is performed from the direction with the higher priority until the feature point is encountered again. Therefore, in the cases of FIGS. 4(b) and 4(C), tracking is performed from the direction with the smallest number. Note that while tracking the contour, the direction is changed according to the rules shown in FIGS. 5(a) to 5(C).

FIG. 8(a) shows the state on the map when all the tracking processing for contour information having feature points in the algorithms ① to ① described above has been completed. In addition, the tracking process for contour information that does not have feature points in the above-mentioned algorithms Tracking is performed from the position on the image data shown in Figure (b) until it returns to the original position.

FIG. 9 is a processing flow for executing the above-described algorithm according to the present invention.

First, in step Sl, target image data is scanned using a 2×2 mask to obtain contour points and feature points, and in step S2, 1 is written for contour points and 2 is written for feature points on the map. .

Then, in step S3, it is determined whether all of the target image data has been scanned, and if the scan has not been completed, the operations from step S to step S2 described above are repeated, and if the scan has been completed, the process moves to step S4. .

In step S4, the map is scanned from the top of the map, and during this scanning, it is determined in step S whether or not the feature point K is found. If the feature point K is found, the feature point is determined in step S6. Contour tracking is started from the position of the image data corresponding to K. Then, in step S, the contour point on the map that has been tracked is erased, and in step Ss, the map is scanned again from the position of the above-mentioned feature point, and this operation is repeated until no new feature point is found. do.

In addition, if it is determined that the feature point K is not found in the above step S, the process moves to step S, where the map is scanned from the beginning of the map,
Step S. It is determined whether or not there is a contour point, and if it is determined that there is no contour point, the process is terminated.

If it is determined in step S1° that there is a contour point, the contour point R found in step S
Start contour tracking from the position of the image data corresponding to .

Then, in step S1□, the contour point on the map that has been tracked is erased, and in step SI3, the map is scanned again from the position of the contour point R described above, and this operation is repeated until no new contour point is found. The process ends when there are no more contour points.

〔effect〕

According to the present invention described above, the positional information of all feature points and contour points is checked by one scan of the target multivalued image, and contour line tracing processing is performed based on the positional information. In particular, contour tracing processing for multivalued images can be performed efficiently and at high speed in a short time.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing a scanning method using a 2×2 mask of image data in the present invention, FIG. 2 is a diagram showing a scanning mode within the mask in the present invention, and FIG. 3 is a diagram showing a map in the present invention. FIG. 4 is a diagram showing the correspondence relationship of image data; FIG. 4 is a diagram for explaining the tracking direction from the feature point in the present invention; FIG. 5' is a diagram showing the direction change rule during contour tracking in the present invention; The figure shows an aspect of the map created according to the present invention. Figure 7 is a diagram for explaining the scanning method on the map and image data according to the present invention. Figure 8 is contour information with feature points according to the present invention. Figure 9 is a flowchart showing the procedure of the contour tracing method according to the present invention; Figure 10 is the conventional contour tracing process. FIG. 11 is a diagram for explaining a conventional method of finding adjacent boundary points. Patent applicant Ricoh Co., Ltd. (dl
(e) (f) ((1)
Figure 1 (a), (b), and (c). (°' (b) (a) (b) (C)
Fig. 4 7・Riff 0 (a) (b) Fig. 6
(bl (C)■
1 ge・■1隨勝・Ki0. ! ,・Shuyato = 4 Yahogetsu Figure 11 Figure 9 = ``&uKEI O:!Toyff'ltW?#
・'Inner i1i'! ¥'? '#Figure 10

Claims (1)

[Claims] A method for performing contour line tracing processing on a multivalued image at high speed, which calculates the position information of all feature points and contour points by one scan of the target multivalued image, and calculates the position of the target multivalued image. Contour tracing is performed sequentially from the positions of the above feature points based on the information,
A contour tracing processing method for a multivalued image, characterized in that after contour tracing from the positions of all of the characteristic points is completed, contour tracing is performed for contours having no characteristic points.