JPH0326876B2 - - Google Patents

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention is an improvement of an image data processing method for identifying an object by using characteristics such as the radius of the image data read by a reading section. Regarding.

In an image data processing system, that is, a processing system that identifies an object in image data read by a reading unit, the feature of the shape of the object is extracted and identification is performed based on this. There is a need to improve processing efficiency when doing so.

[Conventional technology]

A conventional example of a processing system for identifying an object in image data will be explained with reference to the drawings. FIG. 5 is a block diagram illustrating a conventional example.

In FIG. 5, image data I of an image indicated by 2 is read by the reading section 1. The read image data I is divided into a background and an object H by a binarization section 3, and then sent to a contour extraction section 4.
The contour extraction section 4 stores contour point data a, b, c, d, e, . . . , z of the object H in the storage section 5 as shown. The feature extractor 6 extracts features such as the center of gravity, area, and peak distance of the vector radius, that is, the longest value and the shortest value, from the contour point data in the storage part 5, and stores them in the memory 7. The identification unit 8 uses these characteristic parameters to identify the read object.

FIG. 6 shows contour point data extracted from the storage unit 5 of FIG. In other words, the object H in Figure 5
is represented by contour points a to z in FIG.

Conventionally, when finding the longest and shortest values of the vector radius,
For all contour points a to z in FIG. 6,
A method has been adopted in which the radius vectors A to Z are determined and the longest and shortest values are determined from them.

[Problems to be Solved by the Invention] As mentioned above, when calculating the longest and shortest values of the vector radius, in the conventional method, the vector radius A to Z is calculated for all contour points a to z, so the processing is necessary. There was a time-consuming problem.

[Means for solving problems]

The above problem is that the device has a reading section and a processing section, and at least one of the longest distance or the shortest distance among the radial lengths between the plurality of outline points of the image read by the reading section and the center of gravity. In an image processing system that calculates a peak distance of a vector, a selection means for selecting a contour point whose radius vector length is to be calculated, a means for calculating a radius vector length connecting a center of gravity and the selected contour point, and a plurality of calculated peak distances. means for determining the peak distance from the radial length of the contour point, and calculating the radial length for each of the plurality of contour points to obtain a first radial length group; Determine a temporary peak distance from among them, calculate the radius vector length for contour points around the contour point indicating the temporary peak distance to obtain a second radius vector length, and then calculate the second vector radius. This problem is solved by the image data processing method of the present invention, which determines the peak distance from among the radial length group to obtain the peak distance of the vector radius of the image.

[Effect]

As described above, when the number of contour point data exceeds a predetermined value, the present invention calculates the radius of about half of the contour points of the limbal point data, and calculates the longest radius of the radius. and determine the shortest value. Therefore, fewer calculations are required to determine the vector radius than in the conventional method, improving processing efficiency.

〔Example〕

Hereinafter, the present invention will be explained with reference to the drawings. FIG. 1 is a block diagram explaining one embodiment of the present invention, and FIG.
Figure 3 is a data diagram explaining one embodiment of the present invention.
4 and 4 are flowcharts.

In FIG. 1, image data I of an image indicated by 2 is read by a reading unit 1 and separated into a background and an object H by a binarization unit 3, and then a contour extraction unit 4
sent to. In this contour extraction unit 4, the object H
The contour point data a to z are extracted and stored in the storage unit 5.

The outline of the present invention for determining the longest and shortest values of the vector radius from the contour point data a to z is as follows. FIG. 2 is a data diagram showing contour points extracted from the image data I, and an outline of the processing of the processing unit 9 in FIG. 1 will be explained in accordance with the flowchart in FIG. 3.

The processing unit 9 is Determine the number of contour points (a-z).

If the number is 17 or less, the vector radius is calculated for all the contour points as usual (because the method of the present invention requires calculation to obtain a minimum of 17 vector radius).

On the other hand, when the number of contour points exceeds 17,
Among the contour points a to z shown in FIG. 2a, for example, as shown in FIG. 2b, only the odd-numbered contour points, that is, a, c, e...
C, E...Calculate Y. The longest value and the contour points (longest point and second length point) having the longest value and second length value are determined from among the obtained radius vectors. Similarly, the shortest value and the contour points (the shortest point and the next shortest point) having the shortest value and the next shortest value are determined from among the obtained radius vectors.

Calculate the vector radius of the contour points on both sides of the longest point and second longest point obtained in the above section.

Compare the radius vector (calculated value) obtained in the above section with the longest value of the radius vector obtained in the section above.

Calculate the vector radius of the contour points on both sides of the shortest point and the next shortest point obtained in the above section.

The radius vector (calculated value) obtained in the above section is compared with the shortest value of the vector radius obtained in the section above.

Through the above processing, the longest value and the shortest value of the radius vector are determined.

FIG. 4 is a flowchart showing the processing in the above sections 1 to 3 in more detail. In Figure 4, the number of contour points is determined.

When the number of contour points is 17 or less, calculation using the conventional method (calculating the vector radius of all contour points) is performed.

When the number of contour points exceeds 17, calculate the radius vector A of contour point a in Figure 2 b, and initialize it as the temporary longest value and shortest value.
Contour point a is set as the longest point, second longest point, shortest point, and second shortest point.

Set a counter i (not shown) to 1.

Add 2 to the counter value i.

The i-th contour point indicated by the above counter value i (c, e, g, . . . each time i increases)
) is calculated.

Compare the [calculated value] in the above section with the [longest value] in the radius vector calculated so far, and find the [longest value].
When is larger than [calculated value], move on to processing.

When [Longest value] is smaller than [Calculated value], set the longest point so far as the next length point, set the i-th contour point as the longest point, and set [Calculated value] as [Longest value]. , move on to processing.

When [longest value] and [calculated value] are equal, the i-th contour point is set as the second long point again, and the process moves to step .

Compare the [calculated value] in the previous section with the [shortest value] in the radius vector calculated so far, and find the [shortest value].
When is smaller than [calculated value], move on to processing.

When the [shortest value] is larger than the [calculated value], the previous shortest point is set as the next shortest point, and the i-th contour point is set as the shortest point, and the [calculated value] is set as the [shortest value]. Set and move on to processing.

When [shortest value] and [calculated value] are equal, the next short point i-th contour point is set again, and the process moves to step .

When the value of counter i is less than or equal to the total number of contour points,
Repeat the process.

Through the above processing, for example, as shown in FIG. 2b, the radius vector A is determined as the longest value, V as the shortest value, the longest point as a, the second longest point as k, the shortest point as v, and the second shortest point as q.

In Fig. 2 b, contour points b and z on both sides of the longest point a, contour points j and l on both sides of the sub-length point k,
The radius vectors B, Z, J, L, U, W, P, and R are determined for the contour points u and w on both sides of the shortest point v, and the contour points p and r on both sides of the next short point q. The longest value is finally determined by comparing the radius vectors A, B, Z, J, and L, and the shortest value is finally determined by comparing the radius vectors V, U, W, P, and R.

Referring again to FIG. 1, the processing unit 9 stores in the memory 7 the maximum and minimum values of the vector radius obtained by the above processing.

The identification unit 8 identifies the read object H using characteristic parameters such as the area and center of gravity in the memory 7, and the longest and shortest values of the vector radius determined by the above method.

〔Effect of the invention〕

The present invention eliminates the need to calculate radius vectors for all contour points constituting a figure when extracting feature parameters from figure data, thereby improving processing efficiency during figure processing.

[Brief explanation of the drawing]

FIG. 1 is a block diagram explaining one embodiment of the present invention, FIG. 2 is a data diagram explaining one embodiment of the present invention, and FIGS. 3 and 4 are flow diagrams explaining one embodiment of the present invention. Chart, Figure 5 is a block diagram explaining the conventional example, and Figure 6 is a data diagram explaining the conventional example. In the figure, 1 is the reading section, 2 is the image, 3 is the binarization section, and 4 is the contour extraction section. , 5 is a storage unit, 6 is a feature extraction unit, 7 is a memory, 8 is an identification unit, 9 is a processing unit, 1
0 is the main memory, A, B, B, E, J, K, L,
P, Q, R, U, V, W, Y, Z are radius vectors, G is center of gravity, S is area, a, b, c, d, e, f, p,
q, r, u, v, w, y, z indicate contour points.

Claims (1)

[Claims] 1 It has a reading section and a processing section, and calculates the peak distance of at least one of the longest distance or the shortest distance among the radial lengths between multiple contour points of the image read by the reading section and the center of gravity. In an image processing system that
A selection means for selecting a contour point whose radius vector length is to be calculated, a means for calculating a radius vector length connecting the center of gravity and the selected contour point, and a means for determining a peak distance from the plurality of calculated radius vector lengths. Calculating the radial length for each contour point to obtain a first radial length group, determining a temporary peak distance from the calculated first radial length group, Calculate the radial length of the contour points around the contour point indicating the tentative peak distance and calculate the second
1. An image data processing method characterized in that the peak distance of the radial length of the image is obtained by determining the peak distance from a second group of radial vector lengths.