KR20170077994A

KR20170077994A - Efficient Coordinates Map Generation Method for the Camera Image Correction and Rectification

Info

Publication number: KR20170077994A
Application number: KR1020150188061A
Authority: KR
Inventors: 손행선; 이선영; 민경원
Original assignee: 전자부품연구원
Priority date: 2015-12-29
Filing date: 2015-12-29
Publication date: 2017-07-07

Abstract

There is provided an efficient data coordinate map generation method for camera image correction and registration and an image processing method using the coordinate map generated thereby. The image processing method according to an embodiment of the present invention generates a coordinate map, converts the first image into a second image using the generated coordinate map, and the coordinate map converts the first image coordinate and the second image coordinate Expressed in relative relation. This reduces the storage space required for the coordinate map and reduces the use of the bus to reference the coordinate map.

Description

Technical Field [0001] The present invention relates to an efficient coordinate coordinate map generation method for camera image correction and recitation,

The present invention relates to an image processing technique, and more particularly, to a method of generating a reference coordinate map in generating a new image by rearranging pixel values of an input image.

A coordinate map for converting an input image into a new image is generated based on an absolute coordinate. Therefore, a large number of memories are required because the size of the coordinate map is large. As the size of the image increases, the problem becomes more serious.

In particular, when implemented in a semiconductor such as an embedded system or a SoC (System on a Chip), an increase in memory size according to a coordinate map places a burden on the overall system resources, increases the data bus usage rate, do.

In order to reduce the weight of the memory and the load of the bus, a search for a method for reducing the size of the coordinate map is required.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a method of generating a coordinate map with relative coordinates rather than absolute coordinates in order to reduce weight of a memory and load of a bus .

It is another object of the present invention to provide an image processing method for converting an input image into a new image using a coordinate map generated with relative coordinates rather than absolute coordinates.

According to an aspect of the present invention, there is provided an image processing method comprising: generating a coordinate map; And transforming the first image into a second image using the generated coordinate map, wherein the coordinate map is represented by a relative relationship between the first image coordinate and the second image coordinate.

The coordinate map may be expressed by a deviation between the first image coordinate and the second image coordinate.

Also, in the coordinate map, a part may be represented by a relative relationship between the first image coordinate and the second image coordinate, and the rest may be expressed by the first image coordinate.

And the remainder may be coordinates where the x coordinate is less than m and the y coordinate is less than n.

Also, the first image may be an original image, and the second image may be a corrected image.

The corrected image may be a rearranged image.

According to another aspect of the present invention, there is provided an image processing system comprising: a generating unit for generating a coordinate map; And a conversion unit converting the first image into a second image using the generated coordinate map, wherein the coordinate map is represented by a relative relationship between the first image coordinate and the second image coordinate.

As described above, according to the embodiments of the present invention, a coordinate map used when rearranging the positions of pixels of an image is implemented as relative coordinates rather than absolute coordinates, the storage space necessary for the coordinate map is reduced, The use of the bus for referring to the map can be reduced.

1 is a view showing a process of correcting a lens distortion of a single camera,
FIG. 2 is a diagram illustrating a procedure of a stereo camera,
3 is a block diagram of an image processing system according to an embodiment of the present invention,
FIG. 4 is a flowchart illustrating an image processing method according to another embodiment of the present invention.

Hereinafter, the present invention will be described in detail with reference to the drawings.

1. Create coordinate map using relative coordinates

FIG. 1 is a view illustrating a process of correcting a lens distortion of a single camera, and FIG. 2 is a diagram illustrating a procedure of a stereo camera.

Both of these processes change the position of the pixel of the existing image P (x, y) to produce a new image P '(x', y '). What is used here is a coordinate map M (x ', y') in which a new pixel coordinate (x ', y') is designated by the coordinates of an existing pixel.

For example, if the value of the coordinate map at position (x ', y') is M (x ', y') = (256,326) x ', y') becomes the pixel value P (256, 326) at (256, 326) of the existing image.

This relation can be expressed as follows.

P (x ', y') = P (M (x ', y'))

M (x ', y') has coordinates for the existing image as elements. The data size of M (x ', y') is twice the image size (W * H * 2) since M (x ', y') must have both the x coordinate and the y coordinate.

In addition, in the case of the reference camera, the amount of data increases by the number of cameras.

Here, M (x ', y') represents the coordinates (x, y) of the existing image, so that the range of (x, y) is as follows.

0? M (x ')? W-1

0? M (y ')? H-1

If the coordinate map is used as the absolute coordinates of the existing image as described above, the data size of the coordinate map becomes large. For example, in the case of an HD image, coordinates must be expressed up to 1280, so data of 11 bits or more is required.

On the other hand, in the case of image correction or registration, the deviation between the coordinates of the new image matched by the coordinate map and the coordinates of the existing image is not large. That is, the coordinates of the new image are the same as the coordinates of the existing image, or the peripheral coordinates of the existing image.

Therefore, in the embodiment of the present invention, a coordinate map is expressed by relative coordinates rather than absolute coordinates. Expressing the coordinate map with relative coordinates is advantageous in terms of data size, rather than expressing the coordinate map with absolute coordinates.

Absolute coordinates: M (x ', y') = (x, y)

Relative coordinates: M (x ', y') = (x-x ', y-y') = (x, y)

If the coordinate map is represented by the relative coordinates, relative position information is referred to at the coordinates (x ', y') of the new image, so that it can be used more conveniently from the viewpoint of data addressing.

(X'-1 (x ', y')) as the value of the new image P '(x', y ') according to the coordinate map represented by the relative coordinates, , y'-1).

Thus, if the coordinate map is represented by relative coordinates, the size of the coordinate map does not increase exponentially with the size of the input image.

In the case of image correction or registration, the deviation between the coordinates of the new image matched by the coordinate map and the coordinates of the existing image is within 5% of the maximum coordinate.

Therefore, in the case of an HD-class image with coordinates 0 to 1279, 11 bits are required for one coordinate in absolute coordinates, but relative coordinates are 1280 * 0.05 = ± 64 when the deviation is within ± 5% of the maximum coordinate. need.

Therefore, the data size of the coordinate map is reduced by about 36%, thereby increasing the memory efficiency and the data movement amount of the bus, thereby increasing the performance of the product.

Furthermore, if the relative motion is smaller, the amount of data can be adaptively reduced.

2. Image processing system using relative coordinate map

3 is a block diagram of an image processing system according to an embodiment of the present invention. The image processing system according to the embodiment of the present invention performs image processing using a coordinate map represented by relative coordinates.

The image processing includes filtering and other image processing as well as the above-described correction and registration of the image.

3, the image processing system according to the embodiment of the present invention includes a coordinate map generator 110, a coordinate map memory 120, an image input unit 130, an image memory 140, A display unit 150, a video output unit 160, and a bus 170.

The coordinate map generating unit 110 generates a coordinate map expressed by relative coordinates using data that is a reference / reference for generating a coordinate map. The coordinate map memory 120 stores a coordinate map generated by the coordinate map generating unit 110. [

The image input unit 130 receives images from the camera / storage medium / network, and stores the images in the image memory 140.

The image converter 150 refers to the coordinate map stored in the coordinate map memory 120, converts the image stored in the image memory 140, and stores the image in the image memory 140 again. The video output unit 160 outputs the converted video stored in the video memory 140.

The bus 170 is a data / command / data conversion unit between the coordinate map generation unit 110, the coordinate map memory 120, the image input unit 130, the image memory 140, the image conversion unit 150, And provides an address transfer path.

3. Image processing method using relative coordinate map

FIG. 4 is a flowchart illustrating an image processing method according to another embodiment of the present invention. The image processing method according to the embodiment of the present invention also performs image processing using a coordinate map represented by relative coordinates.

As shown in FIG. 4, first, a coordinate map expressed by relative coordinates is generated using data that is a reference / reference for generating a coordinate map (S210).

Next, an image is inputted from the camera / storage medium / network (S220). In step S230, the image input in step S220 is converted by referring to the coordinate map generated in step S210. Thereafter, the converted image is output in step S230 (S240).

4. Variations

Up to now, a method of generating an efficient data coordinate map for camera image correction and registration and a method of performing image processing using the generated coordinate map have been described in detail with a preferred embodiment.

In addition to the above-described embodiments, it is possible to assume a coordinate map that uses both absolute coordinates and relative coordinates.

That is, a part of the coordinate map is expressed in absolute coordinates, and the rest is expressed in relative coordinates. When the relative coordinate size is relatively large, a region having a small coordinate (for example, in a 100 * 100 size image, a 5 * 5 region in the upper left side (a pixel having an x coordinate of 0 to 5 and a y coordinate of 0 to 5 ) Is represented by absolute coordinates, and the rest is expressed by relative coordinates.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, but, on the contrary, It will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the spirit and scope of the present invention.

110: coordinate map generator 120: coordinate map memory
130: image input unit 140: image memory
150: video converter 160: video converter
170: bus

Claims

Generating a coordinate map; And
And converting the first image to the second image using the generated coordinate map,
In the coordinate map,
Wherein the first image coordinate and the second image coordinate are expressed by a relative relationship between the first image coordinate and the second image coordinate.

The method according to claim 1,
In the coordinate map,
Wherein the difference between the first image coordinate and the second image coordinate is expressed by a deviation between the first image coordinate and the second image coordinate.

The method according to claim 1,
In the coordinate map,
A part is represented by a relative relationship between the first image coordinate and the second image coordinate,
And the remainder is expressed by a first image coordinate.

The method of claim 3,
The remainder,
wherein the x coordinate is smaller than m and the y coordinate is smaller than n.

The method of claim 2,
The first image is an original image,
Wherein the second image is a corrected image.

The method of claim 5,
Wherein the corrected image is a rectified image.

A generating unit for generating a coordinate map; And
And a conversion unit for converting the first image into the second image using the generated coordinate map,
In the coordinate map,
Wherein the first image coordinate and the second image coordinate are expressed by a relative relationship between the first image coordinate and the second image coordinate.

The method of claim 7,
In the coordinate map,
Wherein the first image coordinate and the second image coordinate are expressed by a deviation between the first image coordinate and the second image coordinate.