KR101664749B1

KR101664749B1 - Apparatus for enhancing low light level image and method thereof

Info

Publication number: KR101664749B1
Application number: KR1020150157009A
Authority: KR
Inventors: 최진혁; 박태곤
Original assignee: 현대자동차주식회사
Priority date: 2015-11-09
Filing date: 2015-11-09
Publication date: 2016-10-12

Abstract

The present invention relates to a low-illuminance image enhancement apparatus and method thereof, and more particularly, to a low-illuminance image enhancement apparatus and method thereof, which detects a region of interest in a forward image of a vehicle using an LED (Light Emitting Diode) sensor, The present invention provides a low-illuminance image improving apparatus and method which can improve the brightness and color of an image without distortion by improving brightness and color of the image based on the image.
To this end, the present invention provides a low-illuminance image enhancement apparatus comprising: a region of interest detector that detects a preceding vehicle as a region of interest in a forward image of the vehicle; A first pre-processing unit for calculating a first parameter based on the brightness of the ROI detected by the ROI detecting unit; A second pre-processing unit for calculating a second parameter based on a distance between the rear lamps in the ROI detected by the ROI detecting unit; A third pre-processing unit for vertically bisecting the ROI detected by the ROI detecting unit and then calculating a third parameter based on a difference in brightness between the two ROIs; A parameter calculator for calculating a final parameter by summing both the first parameter calculated by the first pre-processor, the second parameter calculated by the second pre-processor, and the third parameter calculated by the third pre-processor, ; And an image enhancement unit for searching for a set value corresponding to a final parameter calculated by the parameter calculation unit based on a table in which set values corresponding to the respective parameters are recorded, thereby improving the image.

Description

[0001] APPARATUS FOR ENHANCING LOW LIGHT LEVEL IMAGE AND METHOD THEREOF [0002]

The present invention relates to a low-illuminance image enhancement apparatus and a method thereof, and more particularly, to a low-illuminance image enhancement apparatus and method thereof that calculates a low-illuminance improvement parameter for a region of interest in a forward image of a vehicle, &Lt; / RTI >

As interest in digital image processing has increased, various techniques, devices, or techniques have been proposed for digital image processing. Such image processing is utilized in various fields as a direction for improving human visual perception ability.

Image extraction, image enhancement, image restoration, image reconstruction, image analysis, image recognition, image compression, and so on. In the image processing that can be classified, techniques matching the above-mentioned various objectivity are used.

A captured image obtained under a limited condition according to physical limitations of the image capturing apparatus for capturing an image or the characteristics of the subject and various external environments in which the subject is present is expressed as a result image according to the marginal situations.

As a method for solving such a problem, there is utilized a method of improving the image, particularly a method of correcting the illuminance and the like of the image, that is, a method for enhancing the perception ability of the human image result by reflecting the actual state of the object , A histogram stretching method, a gamma curve method, a logarithmic transformation method, or a method of adjusting a forced brightness coefficient value.

Among conventional methods, stretching image processing is a technique for adjusting a distribution of pixel values of an image having low contrast to be distributed in a wider area.

In the gamma correction method or the logarithmic conversion method, the original image is used as an independent variable by using a specific function such as an exponential function or a logarithmic function, and the function value is converted into a resultant image by using the function value according to the function as a dependent variable .

However, the above-described method adopts a method of changing the brightness information so that the low-illuminance region becomes high-level by a one-way conversion method while excluding the overall characteristics of the image based on the human perception power of the brightness of the image or the like.

According to this method, although a dark region can be expressed as a bright region, it is possible to reduce the brightness and color of the entire image or to reduce the color saturation of the original image due to distortion of image information, There is a problem that the brightness and color are greatly influenced.

According to an aspect of the present invention, there is provided an image processing apparatus for detecting a region of interest in a forward image of a vehicle and improving the brightness and color of the image based on low- Thereby improving the compensation of the brightness and color of the image without distortion, and a method therefor.

The objects of the present invention are not limited to the above-mentioned objects, and other objects and advantages of the present invention which are not mentioned can be understood by the following description, and will be more clearly understood by the embodiments of the present invention. It will also be readily apparent that the objects and advantages of the invention may be realized and attained by means of the instrumentalities and combinations particularly pointed out in the appended claims.

According to an aspect of the present invention, there is provided an apparatus for enhancing a low-illuminance image, the apparatus comprising: a region of interest detector for detecting a preceding vehicle as a region of interest in a forward image of the vehicle; A first pre-processing unit for calculating a first parameter based on the brightness of the ROI detected by the ROI detecting unit; A second pre-processing unit for calculating a second parameter based on a distance between the rear lamps in the ROI detected by the ROI detecting unit; A third pre-processing unit for vertically bisecting the ROI detected by the ROI detecting unit and then calculating a third parameter based on a difference in brightness between the two ROIs; A parameter calculator for calculating a final parameter by summing both the first parameter calculated by the first pre-processor, the second parameter calculated by the second pre-processor, and the third parameter calculated by the third pre-processor, ; And an image enhancement unit for searching for a set value corresponding to a final parameter calculated by the parameter calculation unit based on a table in which set values corresponding to the respective parameters are recorded, thereby improving the image.

Here, the ROI detecting unit detects a preceding vehicle based on a plurality of LED sensors arranged at a predetermined interval in front of the vehicle, and then detects an area of the preceding vehicle as a ROI in a forward image of the vehicle . At this time, the size of the area of the preceding vehicle is set according to the distance from the vehicle.

The first preprocessing unit may include a first histogram generator for generating a first histogram based on the brightness of the ROI detected by the ROI detector; A first standard deviation calculator for calculating a standard deviation based on the first histogram generated by the first histogram generator; And a first normalizer for normalizing the standard deviation calculated by the first standard deviation calculator.

The second pre-processing unit may include a rear lamp detector for detecting a rear lamp in a region of interest detected by the ROI detecting unit. A distance calculator for calculating a distance between the center of the ROI detected by the ROI detector and each rear lamp, and then calculating a difference between the two distances; And a second normalizer for normalizing the distance difference calculated by the distance calculator. At this time, the rear lamp detector detects a group of pixels having a R (Red) value in a region of interest in the image as a rear lamp.

The third preprocessing unit may further include a third histogram generator for generating a third histogram based on the brightness difference between the two regions after vertically bisecting each ROI detected by the ROI detecting unit. A third standard deviation calculator for calculating a third standard deviation based on the third histogram generated by the third histogram generator; And a third normalizer for normalizing the third standard deviation calculated by the third standard deviation calculator.

According to another aspect of the present invention, there is provided a method for improving a low-illuminance image, the method comprising: detecting a preceding vehicle as a region of interest in a forward image of the vehicle; The first pre-processing unit calculating a first parameter based on the detected brightness of the ROI; Calculating a second parameter based on the distance between the rear lamps in the detected region of interest; Calculating a third parameter based on the difference in brightness between the two regions after the third pre-processing unit bisects the detected region of interest vertically; Calculating a final parameter by summing all of the first parameter, the second parameter and the third parameter; And an image enhancement unit retrieving a set value corresponding to the calculated final parameter based on a table in which a set value corresponding to each parameter is recorded, thereby improving the image.

Here, the detection of the ROI may include detecting a preceding vehicle on the basis of a plurality of LED sensors arranged at a predetermined interval in front of the vehicle, and then detecting the area of the preceding vehicle as a ROI in a forward image of the vehicle . At this time, the size of the area of the preceding vehicle is set according to the distance from the vehicle.

In addition, the first pre-processing step may include generating a first histogram based on the brightness of the detected region of interest; Calculating a standard deviation based on the generated first histogram; And normalizing the calculated standard deviation.

In addition, the second pre-processing step may include detecting a rear lamp in the detected region of interest; Calculating a distance from the center of the detected ROI to each rear lamp, and then calculating a difference between the two distances; And normalizing the calculated distance difference. At this time, the rear lamp detecting step detects a group of pixels having a R (Red) value in a region of interest in the image as a rear lamp.

In addition, the third pre-processing may include vertically bisecting the detected region of interest, and then generating a third histogram based on a difference in brightness between the two regions; Calculating a third standard deviation based on the generated third histogram; And normalizing the calculated third standard deviation.

According to the present invention as described above, a region of interest is detected in a forward image of a vehicle using an LED (Light Emitting Diode) sensor, and brightness and color of the image are calculated based on the low- It is possible to improve the compensation of brightness and color of an image without distortion.

FIG. 1 is a block diagram of an embodiment of a low-illuminance image improving apparatus according to the present invention.
2 is a detailed configuration diagram of an embodiment of a first preprocessing unit according to the present invention,
3 is a detailed configuration diagram of an embodiment of a second preprocessing unit according to the present invention,
4 is a detailed configuration diagram of an embodiment of a third pre-processing unit according to the present invention,
5 is a diagram illustrating an example of a region of interest detected by the region of interest detection unit according to the present invention.
6 is a diagram illustrating an example of a histogram generated by a first preprocessing unit according to the present invention,
7 is a diagram illustrating an example of a rear lamp detected by the second pre-processing unit according to the present invention,
8 is a view illustrating an example of a process of calculating a distance difference between a second preprocessing unit according to the present invention,
FIG. 9 is a flow chart of an embodiment of a low-illuminance image improving method according to the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS The above and other objects, features and advantages of the present invention will become more apparent from the following detailed description of the present invention when taken in conjunction with the accompanying drawings, It can be easily carried out. In the following description, well-known functions or constructions are not described in detail since they would obscure the invention in unnecessary detail. Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram of a low-illuminance image improving apparatus according to an embodiment of the present invention.

1, the low-illuminance image enhancement apparatus according to the present invention includes a region of interest detection unit 10, a first pre-processing unit 20, a second pre-processing unit 30, a third pre-processing unit 40, A calculating unit 50, and an image improving unit 60. [

First, the ROI detecting unit 10 detects ROIs from a forward image of a vehicle input through the camera. That is, the ROI detecting unit 10 includes a plurality of LED sensors (LED sensor modules) arranged to face the front of the vehicle at predetermined intervals to detect (distance, position, etc.) the preceding vehicle being driven.

In addition, the ROI detecting unit 10 detects a region (ROI) of the preceding vehicle on the basis of the information (distance, position) of the preceding vehicle detected by the LED sensor module. At this time, the information sensed by the LED sensor module is matched with the image information input through the camera, so that the area of the preceding vehicle can be found in the image based on the distance and the position of the preceding vehicle, It is preferable that the size of the preceding vehicle is set in accordance with the distance from the vehicle and that the area of interest in which the shape of the preceding vehicle taken in the image is rectangular has a rectangular shape.

The region of interest set as described above is, for example, as shown in FIG. 5A shows the ROI 501 and 502 detected using the LED sensor module at night, and FIG. 5B shows the ROI 503 detected using the LED sensor module during the day.

Next, the first preprocessing unit 20 calculates the first parameter based on the brightness of the ROI detected by the ROI detecting unit 10.

Next, the second preprocessing section 30 calculates the second parameter based on the distance between the rear lamps in the region of interest detected by the region of interest detection section 10. [

Next, the third preprocessing unit 40 vertically bisects the ROI detected by the ROI detecting unit 10, and then calculates the third parameter based on the brightness difference between the two regions. At this time, the third preprocessing unit 40 is divided into two halves so that the rear lamps are included in the respective regions.

Next, the parameter calculating section 50 calculates the parameter by using the first parameter calculated by the first preprocessing section 20, the second parameter calculated by the second preprocessing section 30, and the second parameter calculated by the third preprocessing section 30 And the final parameters are calculated.

Next, the image improving unit 60 includes a table in which a set value (gain value) corresponding to each parameter is recorded, and based on this table, a set value corresponding to the last parameter calculated by the parameter calculating unit 50 To improve the image. That is, the image improving unit 60 improves the brightness and color of the image using the set value.

Hereinafter, the components will be described in detail with reference to FIGS. 2 to 4. FIG.

2 is a detailed configuration diagram of an embodiment of a first preprocessing unit according to the present invention.

2, the first preprocessing unit according to the present invention includes a first histogram generator 21, a first standard deviation calculator 22, and a first normalizer 23.

First, the first histogram generator 21 generates a first histogram based on the brightness of the region of interest detected by the region of interest detection unit 10. [ This is as shown in FIG. In FIG. 6, (a) shows a histogram for the region of interest 501, (b) shows a histogram for the region of interest 502, and (c) shows a histogram for the region of interest 503.

Next, the first standard deviation calculator 22 calculates the standard deviation based on the histogram for each region of interest.

For example, the standard deviation for the histogram for the region of interest '501' is 58.5487, the standard deviation for the histogram for the region of interest '502' is 58.5431 and the standard deviation for the histogram for the region of interest '503 is 40.1191 to be.

Next, the first normalizer 23 normalizes the standard deviation calculated by the first standard deviation calculator 22.

For example, if the threshold is set to 50 and the range is normalized from 0 to 0.5 after setting the range to 30 to 60, the normalization result of the standard deviation for the histogram for the region of interest '501' is 0.4758, The normalization result of the standard deviation for the histogram for the region 502 'is 0.4762, and the normalization result for the standard deviation for the histogram for the region of interest 503 is 0.1687.

3 is a detailed configuration diagram of an embodiment of a second pre-processing unit according to the present invention.

3, the second pre-processing unit according to the present invention includes a rear lamp detector 31, a distance calculator 32, and a second normalizer 33. [

First, the rear lamp detector 31 detects a rear lamp in a region of interest detected by the region-of-interest detecting portion 10. That is, the rear lamp detector 31 detects, as a rear lamp, a group of pixels having an R value based on the R (Red) value in the region of interest in the image. This is because the rear lamp of the vehicle is red, as shown in Fig.

7 (a) shows the rear lamp detected in the region of interest '501', FIG. 7 (b) shows the rear lamp detected in the region of interest '502' Represents a lamp.

Next, the distance calculator 32 calculates the distance from the center of the ROI detected by the ROI detector 10 to each rear lamp, and then calculates the difference between the two distances. That is, the distance calculator 32 calculates the distance from the center to the right rear lamp and the distance from the center to the left rear lamp, and then calculates the difference between the two distances. This is as shown in FIG.

In FIG. 8, (a) is a process of obtaining a distance difference with respect to the rear lamp detected in the region of interest '501', for example, 2.1, and (b) (C) is 47.3762, for example, as a process of obtaining a distance difference with respect to the rear lamp detected in the region of interest '503'. At this time, the reason why the distance difference in (c) is large is because the left rear lamp is not detected and the distance is zero.

Next, the second normalizer 33 normalizes the distance difference calculated by the distance calculator 32. Then,

For example, if the threshold is set to 10 and the range is normalized from 0 to 0.25 after setting the range to 50 to 0, the result of normalizing the distance difference to the rear ramp in the region of interest '501' is 0.2395, The result of normalizing the distance difference with respect to the rear lamp in '502' is 0.2452, and the result of normalizing the distance difference with respect to the rear lamp in the region of interest '501' is 0.0475. At this time, the range is set inversely, and the larger the value, the smaller the normalization result.

4 is a detailed configuration diagram of an embodiment of a third preprocessing unit according to the present invention.

4, the third preprocessing unit according to the present invention includes a third histogram generator 41, a third standard deviation calculator 42, and a third normalizer 43.

First, the third histogram generator 41 vertically bisects each ROI detected by the ROI detecting unit 10, and then generates a third histogram based on the brightness difference between the two ROIs.

Next, the third standard deviation calculator 42 calculates the third standard deviation based on the third histogram generated by the third histogram generator 41. [

For example, the standard deviation between the two regions for the region of interest 501 is 0.7071, the standard deviation between the two regions for the region of interest 502 is 7.0711, and the standard deviation between the two regions for the region of interest 503 is 47.3762 to be.

Next, the third normalizer 43 normalizes the third standard deviation calculated by the third standard deviation calculator 42. Then,

For example, if the threshold is set to 10, and the range is normalized from 0 to 0.25 after setting the range to 50 to 0, the normalization result of the standard deviation between the two regions for the region of interest '501' is 0.2465, 502 'is 0.2146, and the normalization result of the standard deviation between the two regions for the region of interest' 503 'is 0.0131.

Based on the thus-calculated normalization result, the parameter calculation unit 50 calculates a parameter.

That is, 0.4758 + 0.2395 + 0.2465 = 0.9618 is calculated as a parameter for the region of interest 501, and 0.4762 + 0.2425 + 0.2146 = 0.9333 is calculated as a parameter for the region of interest 502. The parameter for the region of interest 503 0.1687 + 0.0475 + 0.0131 = 0.2293.

Therefore, the image enhancement unit 60 refers to a bright image as the parameter is closer to 0, and considering the fact that the closer to the value 1 (the sum of the maximum values of the respective normalization ranges: 0.5 + 0.25 + 0.25) It can be judged whether or not it is. That is, the image enhancement unit 60 may determine the low-illuminance image if the final parameter does not exceed the threshold value.

The image enhancement unit 60 has a table in which a set value (gain value) corresponding to each parameter is recorded based on the parameter.

In the above example, the image enhancement unit 60 determines that the image is a dark image because the parameter for the region of interest 501 is 0.9618, and the parameter for the region of interest 502 is 0.9333, 'Is 0.2293, it is judged as a bright image.

As a result, the image enhancement unit 60 improves the brightness and color of the image using the set value corresponding to the parameter based on the table. At this time, the technique of improving the brightness and color of the image itself is a well-known technology and is not described in detail in the present invention. However, the present invention focuses on obtaining a gain value, which is a key element for improving the brightness and color of an image, and if the gain value is obtained in the above-described manner, the brightness and color of the image can be improved without distortion.

In the present invention, the image enhancement unit 60 may be implemented in a camera.

FIG. 9 is a flow chart of an embodiment of a low-illuminance image improving method according to the present invention.

First, the region of interest detection unit 10 detects the preceding vehicle as a region of interest in the forward image of the vehicle (901).

Thereafter, the first preprocessing unit 20 calculates a first parameter based on the brightness of the ROI detected by the ROI detecting unit 10 (902).

Then, the second preprocessing unit 30 calculates the second parameter based on the distance between the rear lamps in the region of interest detected by the region of interest detection unit 10 (903).

Then, the third preprocessor 40 vertically bisects the region of interest detected by the region of interest detection unit 10, and then calculates a third parameter based on the difference in brightness between the two regions (904).

Thereafter, the parameter calculating section 50 calculates the first parameter calculated by the first preprocessing section 20, the second parameter calculated by the second preprocessing section 30, and the second parameter calculated by the third preprocessing section 40 The final parameters are calculated by summing all the third parameters (905).

Thereafter, the image improving unit 60 searches for a setting value corresponding to the final parameter calculated by the parameter calculating unit 50 based on the table in which the setting value corresponding to each parameter is recorded, (906).

Meanwhile, the method of the present invention as described above can be written in a computer program. And the code and code segments constituting the program can be easily deduced by a computer programmer in the field. In addition, the created program is stored in a computer-readable recording medium (information storage medium), and is read and executed by a computer to implement the method of the present invention. And the recording medium includes all types of recording media readable by a computer.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit or scope of the invention. The present invention is not limited to the drawings.

10: ROI detector
20: First pre-
30: Second pre-
40: Third pre-
50: Parameter calculation unit
60: Image improvement section

Claims

A region of interest detector for detecting a preceding vehicle as a region of interest in a forward image of the vehicle;
A first pre-processing unit for calculating a first parameter based on the brightness of the ROI detected by the ROI detecting unit;
A second pre-processing unit for calculating a second parameter based on a distance between the rear lamps in the ROI detected by the ROI detecting unit;
A third pre-processing unit for vertically bisecting the ROI detected by the ROI detecting unit and then calculating a third parameter based on a difference in brightness between the two ROIs;
A parameter calculator for calculating a final parameter by summing both the first parameter calculated by the first pre-processor, the second parameter calculated by the second pre-processor, and the third parameter calculated by the third pre-processor, ; And
An image enhancement unit for retrieving a set value corresponding to a final parameter calculated by the parameter calculation unit based on a table in which a set value corresponding to each parameter is recorded,
Wherein the low-illuminance image enhancing device includes:

The method according to claim 1,
Wherein the ROI detecting unit comprises:
Wherein the controller detects a preceding vehicle based on a plurality of LED sensors arranged at predetermined intervals in front of the vehicle and then detects an area of the preceding vehicle as a region of interest in a forward image of the vehicle. .

3. The method of claim 2,
The area of the preceding vehicle
And the size of the low-illuminance image is set according to a distance from the vehicle.

The method according to claim 1,
The first pre-
A first histogram generator for generating a first histogram based on the brightness of the ROI detected by the ROI detector;
A first standard deviation calculator for calculating a standard deviation based on the first histogram generated by the first histogram generator; And
A first normalizer for normalizing the standard deviation calculated by the first standard deviation calculator,
Wherein the low-illuminance image enhancing device includes:

The method according to claim 1,
The second pre-
A rear lamp detector for detecting a rear lamp in a region of interest detected by the ROI detecting unit;
A distance calculator for calculating a distance between the center of the ROI detected by the ROI detector and each rear lamp, and then calculating a difference between the two distances; And
A second normalizer for normalizing the distance difference calculated by the distance calculator,
Wherein the low-illuminance image enhancing device includes:

6. The method of claim 5,
The rear lamp detector includes:
Wherein a group of pixels having a value of R (Red) in a region of interest in the image is detected as a rear lamp.

The method according to claim 1,
The third pre-
A third histogram generator for vertically bisecting each ROI detected by the ROI detector and then generating a third histogram based on a brightness difference between the two ROIs;
A third standard deviation calculator for calculating a third standard deviation based on the third histogram generated by the third histogram generator; And
A third normalizer for normalizing the third standard deviation calculated by the third standard deviation calculator,
Wherein the low-illuminance image enhancing device includes:

Detecting a preceding vehicle as a region of interest in a forward image of the vehicle;
The first pre-processing unit calculating a first parameter based on the detected brightness of the ROI;
Calculating a second parameter based on the distance between the rear lamps in the detected region of interest;
Calculating a third parameter based on the difference in brightness between the two regions after the third pre-processing unit bisects the detected region of interest vertically;
Calculating a final parameter by summing all of the first parameter, the second parameter and the third parameter; And
The image enhancement unit retrieves a set value corresponding to the calculated final parameter based on a table in which a set value corresponding to each parameter is recorded,
/ RTI >

9. The method of claim 8,
The method of claim 1,
And detecting a preceding vehicle based on a plurality of LED sensors arranged to face the front of the vehicle at predetermined intervals, and detecting an area of the preceding vehicle as a region of interest in a forward image of the vehicle .

10. The method of claim 9,
The area of the preceding vehicle
And the size is set according to the distance from the vehicle.

9. The method of claim 8,
Wherein the first pre-
Generating a first histogram based on the detected brightness of the ROI;
Calculating a standard deviation based on the generated first histogram; And
Normalizing the calculated standard deviation
/ RTI >

9. The method of claim 8,
Wherein the second pre-
Detecting a rear lamp in the detected region of interest;
Calculating a distance from the center of the detected ROI to each rear lamp, and then calculating a difference between the two distances; And
Normalizing the calculated distance difference
/ RTI >

13. The method of claim 12,
The rear lamp detecting step includes:
Wherein a group of pixels having a value of R (Red) in a region of interest in the image is detected as a rear lamp.

9. The method of claim 8,
The third pre-
Generating a third histogram based on the difference in brightness between the two regions after vertically bisecting the detected region of interest;
Calculating a third standard deviation based on the generated third histogram; And
And normalizing the calculated third standard deviation
/ RTI >