KR101487454B1 - method for parallel processing of LOD image - Google Patents

Abstract

The present invention relates to processing a large image, and more particularly to a method for parallel processing of an LOD image capable of processing a large image by performing LOD image generation processing in a parallel scheme by multiple core processors when the large image is divided into multiple regions to generate an LOD image. The method for parallel processing of an LOD image according to the present invention divides a large image into a plurality of regions by a terminal including a multi code processor to process the large image, converts images with respect to the division regions into an independent LOD images and stores the LOD images in different regions in a data memory by a plurality of core processors, and generates an output image by combining the stored LOD images with each other. In this case, the core processors for generating the LOD image generate the LOD image using different global variables. According to the present invention, the core processors generate large LOD image in a parallel scheme, so that the LOD image is generated at a high rate through an efficient use of a resource included in the terminal, and accordingly a time and costs are reduced.

Description

A method for parallel processing of LOD image

The present invention relates to a large-capacity image processing, and more particularly, to an LOD image generation method in which a LOD image is generated by dividing a large-capacity image into a plurality of regions, and a LOD image generation process is performed in parallel through a plurality of core processors, And a method for parallel processing the LOD image.

Geographical Information System (GIS) is a system that creates and manages maps and geographical information used in printed form using a computer, collects, analyzes, and processes data based on the geographical information obtained, The information system is designed to be applied to.

The geographic information system is divided into land information system, city information system, drawing automation and facility management field, but it is used in a wide range of fields such as land, resource, environment, city, ocean, fishery, military, traffic, communication, .

In addition, the geographic information system provides a large-capacity image to be provided as a 3D image through a user terminal in order to provide information related to the terrain in a more practical form.

At this time, a user terminal, for example, a general computer, uses a LOD (Level Of Detail) based image processing method to more easily process a large-capacity image such as a 3D image provided from a geographic information system.

Here, the "LOD" is a method of dividing one image into several levels of resolution so as to smoothly manage a large-capacity image such as a satellite image or an aerial image in the system, and then displaying a proper resolution image according to a user's view position .

 That is, a computer converts a large-capacity image into an LOD-based image using a CPU, and visualizes the image on a display through a graphic card or the like.

Meanwhile, as the amount of information provided through a computer increases and the capacity of information increases, computers currently used in general use a multi core processor, that is, a plurality of independent cores The processor is configured as one CPU.

However, in order to process a large capacity image in a computer, a single core processor is allocated and LOD image processing is performed. Thus, there is a problem that a processing speed is somewhat delayed in processing a large capacity image of a certain size or more.

In particular, most local governments and state agencies periodically acquire large-scale image information and are making maintenance and updating work for them. Therefore, a more efficient processing method for a large-capacity image is required in various fields for processing a large-capacity image.

[Patent Literature]

1. Korean Patent No. 1236980 entitled " Apparatus and method for converting image data in a rendering system "

2. Korean Registered Patent No. 0434178 (entitled " Adjustment Method and Adjustment System for Accuracy of Image Data Interlocked with Video Signal Processing Capability of Terminal means)

Accordingly, it is an object of the present invention to provide a computer apparatus based on a multicore processor, in which a large capacity image is converted into an LOD image and a LOD image is simultaneously generated in parallel using a plurality of core processors, There is a technical purpose of providing an LOD image parallel processing method which enables high speed LOD image processing.

According to an aspect of the present invention, there is provided a method for parallel processing an LOD image in a terminal for processing a large capacity image, the method including dividing a large capacity image into a plurality of regions, And a plurality of core processors, respectively, for outputting the divided images provided in the first step to an LOD image composed of a plurality of tile images of a predetermined size A third step of storing the LOD image generated by the corresponding core processor in a predetermined area of the data memory, and a third step of combining the LOD images stored in the data memory to generate an output image, Wherein the plurality of core processors capable of parallel processing in the first step comprise: Wherein the number of core processors included in the terminal is set to be equal to or less than the number of core processors provided in the terminal, and the partition area of the large capacity image is set to be equal to or greater than the number of core processors capable of parallel processing.

delete

delete

delete

delete

delete

Also, in the second step, each core processor capable of parallel processing performs an LOD generation process using a predetermined global variable.

In the third step, each of the core processors capable of parallel processing sets coordinate values corresponding to positions of a plurality of tile images of the LOD area allocated to the core processor, sets a storage folder name based on the coordinate values, The tile image corresponding to the tile image is stored in the corresponding storage folder.

According to the present invention, in processing a large capacity image in a terminal having a multicore processor, it is possible to generate an LOD image in parallel by using a large number of images through a plurality of core processors, Speed LOD image generation process, as well as a time and cost reduction effect.

In addition, when parallel processing is performed through the plurality of core processors, global variables are set in each core processor to process the LOD image, so that interference between the core processors does not occur, and more efficient LOD image processing becomes possible.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing an internal configuration of an LOD image parallel processing apparatus according to the present invention, which is functionally separated; FIG.
Fig. 2 is a block diagram showing the function of the configuration means 300 shown in Fig. 1; Fig.
FIGS. 3 to 5 are diagrams illustrating an LOD image processing process performed by the LOD processing unit 400 shown in FIG. 1; FIG.
FIG. 6 is a flow chart for explaining the operation of the LOD image parallel processing apparatus shown in FIG. 1. FIG.
FIG. 7 is a graph showing an execution time measurement result of the LOD image processing execution time according to the number of core processors in the LOD image parallel processing apparatus shown in FIG.

Hereinafter, a preferred embodiment of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is a diagram showing a schematic configuration of an LOD image parallel processing apparatus according to the present invention.

1, the LOD image parallel processing apparatus according to the present invention includes an image analysis processing unit 100, an image segmentation processing unit 200, an environment setting unit 300, an LOD processing unit 400, And an image generating means (500). At this time, the LOD processing unit 400 includes a plurality of core processors 410 and a data memory 420.

The image analysis processing means 100 extracts various coordinate information constituting the spatial region information including the coordinate system, coordinates, and earth ellipsoid of the corresponding image from the inputted large-capacity image.

The image segmentation processing means 200 divides a large-capacity image into a plurality of regions and provides an image for each divided region to the LOD processing means 400 in a parallel manner. At this time, the image segmentation processing unit 200 divides the large capacity image into a number of image areas corresponding to the number of parallel processing core processors constituting the LOD processing unit 400, and provides each divided image to each core processor in a parallel manner do.

The environment setting unit 300 sets a resolution level for the large-capacity image and sets a core processor 410 capable of parallel processing at present. The environment setting unit 300 may set different global variables for the parallel processor 410 and store the LOD image processed in each core processor 410 in a different memory Area. In addition, the environment setting unit 300 sets the number of divisions of the LOD image corresponding to the plurality of core processors 410 capable of parallel processing, and sets the number of tiles and the size of the tiles of the LOD image to correspond to the respective resolution levels.

The core processor 410 configured in parallel by the LOD processing means 400 tiles the corresponding divided image provided from the image segmentation processing means 200 to convert it into an LOD image composed of a plurality of tiles, 420).

The output image generating means 500 reads the LOD image stored in the data memory 420 of the LOD processing means 400 and combines them to generate an output image. At this time, the output image generation means 500 reads the LOD images processed by the core processors 410 in the order stored in the corresponding storage region of the data memory 420, and combines the LOD images of the different image regions , And an output image is generated.

FIG. 2 is a functional diagram showing the internal configuration of the configuration means 300 shown in FIG.

The environment setting unit 300 includes a resolution level setting block 310, a parallel processing setting block 320, and an LOD environment setting block 330 as shown in FIG.

The resolution level setting block 310 sets at least one resolution level according to the user's position with respect to the large-capacity image. For example, the resolution level setting block 210 may set the resolution level to a level of 0 to 10 according to the resolution. At this time, the "10" level is set as an original large capacity image, for example, an LOD having 1,000 tiles, a "9" level is an LOD having 500 tiles, , ... , ≪ / RTI >

The parallel processing setting block 320 includes a number of core processors that can be parallel-processed among a plurality of core processors 410 provided in the computer device, Allocate a memory area. In this case, when the computer device includes four core processors 410, the parallel processing setting block 320 sets all of the four core processors to LOD parallel processing according to a work environment such as a computer use state of the user , Or only three core processors can be configured to perform LOD parallel processing. The setting of the LOD parallel processing core processor 410 may be automatically set at the corresponding computer device or may be configured to be set by the user at will. In addition, the parallel processing setting block 320 allocates a global variable to the LOD parallel processing core processor 410. This makes it possible to allocate different variables to each of the parallel processing core processors 410 and prevent the core processor 410 from using the same variable in the other core processors 410 for the variables assigned to the arbitrary core processors 410. [ So as to prevent interference between them.

), 각 병렬 처리 코어 프로세서(410)에 할당되는 타일의 수를 설정한다. 예컨대, 해상도 레벨이 "10"레벨인 영상에 대해 1,000개의 타일수가 설정되고, 4개의 코어 프로세서(410)가 병렬처리 가능한 경우, 10레벨 영상에 대해 제1 내지 제4의 LOD 영상으로 분리되고, 각 LOD 영상은 250개의 타일을 갖도록 설정된다.In addition, the LOD configuration block 330 sets the number of divisions of the LOD image so as to correspond to a plurality of core processors 410 capable of parallel processing, and sets the number of tiles for each image and the size . At this time, the LOD setting block 230 divides the number (M) of all tiles corresponding to the resolution level for each resolution level by the number (N) of parallel processing core processors

), And sets the number of tiles allocated to each parallel processing core processor 410. For example, if 1,000 tiles are set for an image having a resolution level of "10 " and 4 core processors 410 are capable of parallel processing, the 10-level image is divided into first to fourth LOD images, Each LOD image is set to have 250 tiles.

3 to 5 are views illustrating an LOD image processing process of the LOD processing unit 400 shown in FIG.

FIG. 3 is a diagram illustrating a case where there are four core processors 410 capable of parallel processing. In FIG. 3, four divided areas are displayed for one image screen, that is, for example, ①, ②, ③ and ④ The first to fourth divided areas are shown. That is, the first to fourth core processors independently generate the LOD image for the first to fourth segment areas (1, 2, 3, 4).

As shown in FIG. 3, each of the core processors 410 includes first through fourth global variables for the first through fourth divided areas (1, 2, 3, 4), for example, localImage1, localImage2, localImage3 , and localImage4 to perform LOD image processing. At this time, each core processor 410 generates an LOD image corresponding to each level corresponding to all predetermined resolution levels (0 to n).

At this time, the first core processor 410, which generates the LOD image for the core processor 410, for example, the first partition area (1), inserts the image of the first partition area (1) And divides the image into a tile image (T) of a predetermined size and number. For example, the size of each tile image T may be set to 256 x 256 pixels in width and length, and one LOD image may be set to 16 tile images T in four lines each in the horizontal and vertical directions .

In addition, the core processor 410 stores tiling information corresponding to each LOD image, that is, a plurality of tile images T in a predetermined allocation area of the data memory 420. At this time, the core processor 410 assigns a coordinate value corresponding to the image position sequence (index) in the horizontal direction and the vertical direction to the tile image T as a unique ID. That is, each tile image coordinate value is set to "(horizontal direction index, vertical direction index) ". In addition, a storage folder for storing a corresponding LOD image is generated based on the horizontal index or the vertical index, and the corresponding tile image is stored in the storage folder. For example, when a storage folder is created on the basis of the vertical index, a storage folder named "0000" is created by extending the line with the vertical index "0 " (0,0), (1,0), (2,0), (3,0), and (0,0) of the tile coordinates for the folder name "0000 " Similarly, tile images having tile coordinates with a vertical direction index of "0" are stored in different "0000" folders as different image files. This is because it is not known how many core processors 410 capable of parallel processing perform LOD image generation processing independently of each other and at which time processing by the core processor 410 is performed first, To clarify the location of the LOD image in the large-capacity image.

FIG. 5 illustrates a storage folder structure formed in the data memory 420. As shown in FIG. 5, the data memory 420 sets a tile image index (0000 to xxxx, P2) for each resolution level (0 to n, P1) in a preset memory area, The tile image T is stored in the form of an image file (0000_0000.jpg, 0000_0001.jpg, 0000_0002.jpg, 0000_0003.jpg, P3).

FIG. 6 is a flowchart illustrating an operation of the LOD image parallel processing apparatus shown in FIG. 1. Referring to FIG.

First, basic information for LOD image processing is set in the configuration unit 300 of the LOD image parallel processing device. That is, at least one resolution level for a large-capacity image is set, and the total number of tiles for the image of one frame and the size information of each tile are set in advance as basic information. At this time, the resolution level, the total number of tiles, and the size of the tile may be automatically set to be optimized by the LOD image parallel processing device or arbitrarily set by the user.

In the above state, the LOD image parallel processing apparatus sets up a parallel processing environment for converting a large capacity image into an LOD image by a parallel processing method (ST1). That is, the configuration unit 300 of the LOD image parallel processing apparatus confirms the number of parallel processors capable of parallel processing in order to generate an LOD image in the current state of the terminal, and sets different global variables for each core processor 410 And allocates different memory areas in the data memory 420 to each core processor 410. In addition,

The image segmentation processing means 200 sets the image segmentation information corresponding to the number of parallel processing core processors 410 provided from the environment setting means 300, that is, the number of LODs and the divided region coordinate information. At this time, the image segmentation processing unit 200 can set the LOD segmentation area to be equal to or greater than the number of the core processors 410. [ This is a case in which one core processor 410 can process two or more LODs.

Then, when a large-capacity image is input to the image analysis processing means 100 (ST2), the image analysis processing means 100 extracts various coordinate information including spatial region information on the inputted large-capacity image, Together with the spatial coordinate information, to the image segmentation processing means 200.

The image segmentation processing unit 200 divides the large capacity image into regions corresponding to the predetermined number of LODs based on the spatial coordinate information and the divided region coordinate information, and provides the partitioned image to a core processor capable of parallel processing (ST3).

 Each core processor 410 of the LOD processing means 400 generates an LOD image based on the divided image provided from the image segmentation processing means 200 (ST4). At this time, the plurality of core processors 410 capable of parallel processing independently generate LOD images for corresponding divided images using different global variables.

Next, each core processor 410 stores the LOD images independently generated in a predetermined area of the data memory 420 (ST5). At this time, each core processor 410 divides the LOD image so as to correspond to a preset number of tiles corresponding to the current resolution level, and stores it in the data memory 420. In addition, each of the core processors 410 sets a coordinate value corresponding to a position of the multi-tile image of the LOD area allocated to the core processor 410, sets a storage folder name based on the coordinate information, Folder.

The LOD processing unit 400 stores the LOD image generated in parallel in the plurality of core processors 410 in the respective areas of the data memory 420 so that the output image generating unit 500 generates The tile image of the area, that is, the image file is read and combined by a method corresponding to the stored method, thereby generating and outputting the output image (ST6). That is, the output image generating unit 500 reads the tile images constituting each LOD image corresponding to the resolution level set by the user, combines the tile images to be arranged at the positions corresponding to the LOD divided areas, and generates an output image .

However, when two or more multiple resolution levels are set, the LOD images for the respective resolution levels may be stored in different data memories 420, And provide an output image for the resolution level. At this time, the generation of the LOD image is processed for each resolution level, thereby preventing an error such as an image overlap between the upper level and the lower level.

Meanwhile, FIG. 7 is a graph illustrating the execution time of the LOD image processing according to the number of core processors in the LOD image parallel processing apparatus shown in FIG.

As shown in FIG. 7, when the execution time of the LOD image processing of the same large capacity image is measured, it is about 700 seconds when the number of core processors is 1, about 400 seconds when the number of core processors is 2, , And about 290 seconds in the case of 8, the time required to process the LOD image by the parallel processing method using four core processors is twice or more than that of the LOD image processing using the single core processor Able to know. At this time, the execution time is a time measured by adding the sequential processing time including the division of the area of the large capacity image, and the LOD processing time according to the number of core processors, by analyzing the coordinates of the large capacity image.

That is, according to the embodiment, in processing a large capacity image in a terminal equipped with a multicore processor, a large capacity image is divided into regions corresponding to a plurality of core processors, and LOD The LOD image generation process can be performed faster and faster by efficiently using the resources provided in the terminal.

In the present invention, when parallel processing is performed through the plurality of core processors, global variables are set in each core processor to process the LOD image, so that no interference occurs between the core processors.

Although the present invention has been described in connection with the above-mentioned preferred embodiments, it is possible to make various modifications and variations without departing from the spirit and scope of the invention. Accordingly, the scope of the appended claims should include all such modifications and changes as fall within the scope of the present invention.

100: image analysis processing means, 200: image division processing means,
300: environment setting means, 400: LOD processing means,
500: output image generating means, 310: resolution level setting block,
320: parallel processing setting block, 330: LOD configuration block,
410: core processor, 420: data memory.

Claims (8)

delete delete delete delete A method for parallel processing an LOD image in a terminal configured to have a multicore processor and process a large capacity image,
A first step of dividing a large capacity image into a plurality of areas and outputting images for each of the divided areas in parallel form to a plurality of core processors,
The plurality of core processors may include a second step of converting the divided image provided in the first step into an LOD image composed of a plurality of tile images of a predetermined size,
Each of the core processors storing a LOD image generated by the core processor in a predetermined area of the data memory,
And a fourth step of combining the LOD images stored in the data memory to generate an output image,
In the first step, the plurality of core processors capable of parallel processing are set to be equal to or less than the number of core processors provided in the terminal, and the divided area of the large capacity image is set to be equal to or larger than the number of core processors capable of parallel processing To-parallel processing method. delete 6. The method of claim 5,
Wherein each core processor capable of parallel processing in the second step performs an LOD generation process using a predetermined global variable. 6. The method of claim 5,
Each of the core processors capable of parallel processing in the third step sets a coordinate value corresponding to a position of a multi-tile image of an LOD area allocated to the core processor, sets a storage folder name based on the coordinate value, And the tile image is stored in the corresponding storage folder.

Images