JPH02257230A - Managing method for two-dimensional image memory - Google Patents

Abstract

PURPOSE:To avoid waste from on an image memory and to efficiently store the images of the maximum quantity by storing plural images in the image memory, and using an image memory space as the combination of divided rectangular areas of a prescribed size in accordance with the size of each storing image. CONSTITUTION:The minimum unit area of length LX/(N-th power of 2 (N is an integer of >='0')) and LY/(N-th power of 2) in the X and the Y directions is divided into rectangular areas called as sectors 100a, and numbers are applied as sectors '0', 1, 2, - 63 successively from the position near the origin of two X and Y axes. Based on the sector being the minimum unit of this rectangular area, the storage of an image to an image memory 100 is managed, and an area constituted of sectors of pieces of M-th power of 4 (M is an integer of >='0' and <N) for forming adjacent rectangular areas is set as the cluster of a type M. When 4 pieces of clusters of the type M are accumulated, only one piece of cluster of a type (M + 1) being its high-order type is generated, and on the contrary, when one piece of cluster of the type (M + 1) is decomposed, 4 pieces of adjacent clusters of the type M are generated. In such a manner, a memory area which becomes useless can be decreased.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention stores image data of a plurality of images having arbitrary lengths in the X and Y directions in an image memory to which two-dimensional addresses are assigned. This invention relates to a method for managing two-dimensional image memory required when

[Conventional technology]

Conventionally, with this type of means, when it is necessary to store image data of multiple images in an image memory to which two-dimensional addresses are assigned, the maximum X and Y of the stored images are
The two-dimensional space of the image memory is uniformly divided into rectangular areas with a length in the direction, and the stored image is
The data is stored in one of the free areas.

[Problem to be solved by the invention]

By the way, in this conventional method, the image memory area of the same capacity is used regardless of the length of the image in the X and Y directions, so if the stored image is smaller than the divided rectangular area, it will be wasted. The disadvantage is that the image memory area increases.

In view of the above-mentioned problems in the conventional method, the present invention secures an area with an appropriate memory capacity depending on the length of the stored image in the X and Y directions, and stores the image there. The purpose is to provide a method that makes it possible to reduce wasted memory area.

[Means to solve the problem]

In order to achieve the above object, the present invention Length LX in the X and Y directions of two orthogonal X and Y axes.

The image memory to which the two-dimensional address of LY is assigned is defined as the length LX/(power of 2) in the X and Y directions, and LY/(2
A power of 4) is divided into rectangular areas called sectors, and an area called a cluster is defined which is made up of a power of 4 sectors that are adjacent to each other to form a rectangular area. A cluster allocation table, which is an array for recording the usage status, is provided on the program memory, and images to be stored in the image memory are stored in the smallest cluster that can accommodate them, and images are stored in the image memory. and when performing deletion, update the cluster allocation table and use any X
, in the image memory space by sequentially storing new images starting from the cluster that is closer to the origin of the two axes X and Y, corresponding to the image data of multiple images having lengths in the Y direction. This is a two-dimensional image memory management method characterized in that a plurality of images are arbitrarily stored and deleted.

[For production]

As described above, the combination of lean rectangular image memory areas corresponding to the arbitrary size of the stored image is stored in the optimal area in a certain order, and the image of the largest number of images is stored in the optimal area. Data can be stored and two-dimensional image memory can be managed efficiently.

〔Example〕

An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a sector number diagram showing designation numbers attached to each sector formed by dividing the image memory into rectangular sections.

Let us consider a state in which two-dimensional addresses with lengths LX and LY in the X and Y directions are assigned on the image memory 100, assuming two orthogonal X and Y axes.

Its length in the x and y directions is LX/(2 to the N power).

The minimum unit area of LY/ (2 to the N power) is sector 100a
It is divided into rectangular areas called 0. Sector 1° Sector 2.・・・・・・
・Attach a sector number such as , sector 63.

Here, N is an integer greater than or equal to 0.

Storage of images in the image memory 100 will be managed based on sectors as the minimum units of this rectangular area.

An area composed of 4 M sectors adjacent to each other forming a rectangular area will be referred to as a type M cluster.

Here, M is an integer greater than or equal to 0 and less than N.

When four clusters of type M that form adjacent rectangular areas are accumulated, only one cluster of type (M+ 1 ), which is a higher type, is generated.

Conversely, when one cluster of type (M+1) is decomposed, four adjacent clusters of type M are generated.

Each of these four adjacent clusters of type M is assigned the same number as the number of the first sector, which is the smallest sector of each sector number shown in FIG.

Each cluster has a specific sector as its first sector depending on its type.

FIG. 2 is an explanatory diagram of a cluster area constituted by 4 to the M power of sectors forming adjacent rectangular areas in this manner.

Each image stored in image memory 100 is stored in a cluster of the smallest size it can be stored.

An image stored in a type M cluster is called a type M image.

A cluster in which an image is stored is referred to as an in-use cluster, and a cluster in which an image is not stored is referred to as an unused cluster.

In order to store the arrangement of clusters in the image memory 100 and their usage status, a cluster allocation table, which is arranged in order of sector status and image number, is created.
Provided on program memory. The form of this cluster allocation table is shown in Figures 4 to 1 of the drawings, which will be described later.
This is shown in (c) of each figure.

Each time an image is stored or deleted, the sector status and image number in the cluster allocation table are updated.

The sector status of the first sector of an unused cluster is set to "the value of the cluster type plus 1", and the sector status of the first sector of a used cluster is set to "the value of the cluster type plus 1". The value of adding 1" [The reason why 1 is added to both in this way is to identify it because an unused sector is represented by a 0 code as described later. Set the value.

Furthermore, the [identification] number of the image to be stored is set in the image number.

In addition, the sector status of the middle sector of each cluster is set to 0.

Here, the operation of the calculation method for storing an image in the image memory of the present invention will be explained.

When storing images, perform the following processing.

The type T of the image is determined from la and the length of the image in the X and Y directions.

2a. Check whether there is an unused cluster of type T. If it does not exist, an unused cluster of a type higher than type T is searched, and the one of the lowest type and the lowest number is decomposed to generate an unused cluster of type T. If there is no unused cluster of a higher type than type T, the process is interrupted because no image can be stored.

3a, Store the image in the lowest numbered cluster among the unused clusters in 2a.

When erasing an image, perform the following processing.

tb, the used cluster A in which the image was stored is set as an unused cluster, and unused clusters in its vicinity are accumulated as many times as possible to generate an unused cluster B.

2b. Check whether there is a cluster in use that is of the same type as unused cluster B and has a higher number than unused cluster B. If there is no cluster with a higher number than unused cluster B, the process ends.

3b, among the clusters in use in the previous 2b, set the cluster C with the highest number as an unused cluster, transfer the image to the unused cluster B, change cluster B to the used cluster B, and then transfer it from the previous lb. The previous processing is repeated for cluster C, which has the largest number.

Next, a cluster definition diagram of this embodiment is shown in FIG.

In this embodiment, the number of image memories is 4 to the power of 3, or 6
It is divided into four sectors [Fig. 3(b)], and clusters from type 0 to type 3 [Fig. 3(c) to (f)] are defined.

From now on, we will describe a management method for storing each image in this two-dimensional image memory, assuming cases from the first stage to the ninth stage.

In the initial state of the first stage, unused cluster 0 of type 3 is arranged.

The state is shown in FIG. 4, where FIG. 4(a) is a state diagram of the status in the sector, FIG. 4(b) is a diagram of the arrangement of image numbers of images to be stored, and FIG. 4(c) is a diagram of the program. Cluster φ allocation table represented in memory.

Hereinafter, each of (a) and (
The same applies to b) and (c).

In FIG. 4(a), the sector numbered 0 of an unused cluster of type 3 is set to 4, which is the sum of the cluster type value [-3] and 1, and all the rest are set to 0.

In FIG. 4(b), all values are 0 because no image has been given yet.

Figure 4 (C) shows sector 0° sector 1 in the sector number column.
．． ~． The sector status and image number in the sector 63 are as shown in FIGS. 4(b) and 4(c), respectively.

When storing image number 1 of the second stage type 0 in the image memory,
Since there are no unused clusters of type 0, type 3
Cluster 0 is decomposed into four type 2 clusters [each cluster is assigned the same number as the first lowest sector number, and cluster 0. cluster 16
．． cluster 32, cluster 48], and then decomposes the lowest numbered cluster 0 to create four type 1 clusters [type 1 cluster 0.4.8. 12]
, and further decomposes the lowest numbered cluster 0 to create 4 type 0 clusters [type 0 cluster 0. 1. 2. 3], and here, image number 1 is stored in cluster 0, which is the lowest number among them.

This aspect is represented in FIG.

That is, image number 1 is stored in sector 0, which is the first sector of the cluster in use.
“Value of cluster type value [-〇] plus 1 [1]
A value [-1] is set by inverting the sign of `` (that is, a negative sign).For other unused clusters, the same applies to FIG. 4.

3rd stage Subsequently, when storing image number 2 of type 0, the 6th stage
As shown in the figure, image number 2 is stored in cluster 1, which is the lowest number among the unused clusters of type 0.

Fourth stage Furthermore, as shown in FIG. 7, image number 3 of type 0,
When storing 4.5, after storing image number 3.4 in cluster 2.3, which is a type 0 cluster, as in the third stage, there will be no unused type 0 cluster, so Decomposing type 1 cluster 4, type 0
Four clusters are created, and image number 5 is stored in cluster 4.

Fifth step: Subsequently, when storing type 1 image number 6.7.8, image number 6.7 is stored in type 1 unused cluster 8, 1.
2, there will be no unused type 1 clusters, so the type 2 cluster 16 will be decomposed to generate four type 1 clusters, of which cluster 1
Image number 8 is stored in 6. This aspect is illustrated in FIG.

6th stage Furthermore, type O image number 9. 10. 11°12
When storing the image number 9, as shown in Figure 9,
10.11 to cluster 5.6. 7, there are no unused clusters of type 0, so type 1
The cluster 20 of is decomposed to generate four clusters of type 0, and image number 12 is stored in cluster 20 of them.

7th Step Subsequently, when storing image number 2 of type 0, cluster 1 in which it is stored is first set as an unused cluster. Cluster 1 cannot be accumulated any further.

Next, among clusters of type 0, the contents of cluster 20 with the largest number [that is, image number 12] are transferred to cluster 1.
, and makes cluster 1 a busy cluster again. Then, with cluster 20 as an unused cluster, clusters 20, 21, 22, and 23 are accumulated to generate a type 1 cluster. Since there is no type 1 cluster with a number greater than 20, the process ends. FIG. 10 illustrates this process.

8th stage Furthermore, when storing image numbers 13.14 and 15°16 of type 1, image numbers 13°14.15 are stored in clusters 20 and 24.28, as shown in FIG. Later, since there will be no unused type 1 clusters, the type 2 clusters 32 will be decomposed and type 1 clusters will be disassembled.
4 clusters are generated, and image number 16 is stored in cluster 32.

Ninth stage: When erasing type 1 image number 14, first the cluster 24 in which it is stored is made unused. Clusters 24 cannot be accumulated any further. Next, the contents of the cluster 32 with the largest number among the clusters of type 1 (ie, image number 16) are transferred to the cluster 24, making the cluster 24 a used cluster and the cluster 32 an unused cluster. cluster 32
Accumulating ゜36.40.44, type 2 cluster 3
Generate 2. These conditions are represented in FIG.

As is clear from the above description, an image of a suitable size can be placed preferentially to an unused cluster with a small number.

[Effects of the Invention] As explained above, the present invention stores a plurality of images in the image memory, so the image memory space is divided into rectangular areas of a fixed size according to the size of each stored image. Since they are used in combination, it is possible to efficiently store the maximum number of images without any waste on the image memory to which two-dimensional addresses are assigned.For example, if applied to robot vision, template images can be stored. Since the information is efficiently held in the memory, the reliability of template number matching can be significantly improved, which is a significant effect.

[Brief explanation of drawings]

FIG. 1 is a sector number diagram in one embodiment of the present invention, and FIG.
The figure is an explanatory diagram of a cluster area composed of 4 M sectors contiguously forming a rectangular area, FIG. 3 is a cluster definition diagram of one embodiment, and FIGS. This is an explanatory diagram assuming an arbitrary case when storing an image. (a) shows the sector status, (b
) shows an image number allocation diagram, and (C) shows a cluster allocation table in the program memory. 100... Image memory 100a... Sector 100n forming the minimum unit rectangular area
...Sector number corresponding to the position.

Claims (1)

[Claims] 1. Lengths LX and LY in the X and Y directions of two orthogonal X and Y axes
The image memory to which the two-dimensional addresses of Defines an area called a cluster, which is made up of a power of 4 number of sectors that form a rectangular area, and programs a cluster allocation table, which is an array for recording the arrangement of clusters in the image memory and their usage status. An image to be stored in the image memory is stored in a cluster of the minimum size that can accommodate it, and when storing and erasing images in the image memory, the cluster allocation table is updated and any X
, in the image memory space by sequentially storing new images starting from the cluster that is closer to the origin of the two axes X and Y, corresponding to the image data of multiple images having lengths in the Y direction. A method for managing a two-dimensional image memory, characterized in that a plurality of images are arbitrarily stored and deleted.