KR102091521B1 - Apparatus and method for ramp distribution based image processing - Google Patents

Abstract

The present invention relates to an image processing apparatus and an image processing method. An image processing apparatus based on a gradient distribution according to the present invention includes: an image input unit receiving an image; An input histogram calculator calculating a histogram of the received image; A histogram sorting unit to sort the histogram using a sorting algorithm; An inclination distribution calculator for calculating an inclination distribution for the sorted histogram; A target histogram calculation unit calculating a target histogram through the calculated gradient distribution; And a contrast ratio improvement unit for re-sorting the target histogram to calculate a transfer function, and outputting an image with an improved contrast ratio using the transfer function.

Description

Gradient distribution-based image processing device and image processing method {APPARATUS AND METHOD FOR RAMP DISTRIBUTION BASED IMAGE PROCESSING}

The present invention relates to an image processing apparatus and an image processing method, and more particularly, to a method for processing a contrast ratio of an infrared thermal imaging system.

 In general, an infrared thermal imaging device converts heat emitted from the surrounding environment of an object into an electronic signal through an infrared sensor, thereby forming a screen that can be seen by a human eye. This infrared thermal imaging device, unlike a visible light sensor (CCD: Charge Coupled Device), which can acquire images only during the day, can acquire images even at night, and can easily detect targets with a lot of heat. It can be used for military purposes such as tank / armor sighting. In addition, recently, it is also used as a driving assistance device that pre-detects and alerts pedestrians and animals when driving a car at night.

1 is a block diagram illustrating an embodiment of a conventional infrared thermal imaging device.

Referring to FIG. 1, a typical infrared thermal imaging device includes an infrared detection unit 110 that detects infrared rays emitted from an arbitrary object and detects an infrared image signal of 12 to 14 bits, and a non-uniformity of the image signal detected by the infrared detection unit It is composed of an image signal processor 120 that performs correction, noise reduction, contrast enhancement, and the like, and an image output unit 130 that outputs an image signal processed by the image signal processor 120.

2 is a view for explaining an embodiment of the infrared image generated by the conventional infrared thermal imaging device according to FIG.

3 is a graph showing the histogram of FIG. 2.

Referring to FIG. 2, the infrared image generated by the infrared detection unit 110 looks blurry than the visible light image, and infrared rays emitted from most of the long distance (green arrow in FIG. 2) have similar pixel values in the infrared detection sensor.

This is due to the property that heat emitted from an object is easily absorbed, scattered or reflected by moisture particles while passing through the atmosphere. Therefore, the histograms of the 12-bit to 14-bit infrared images obtained from the infrared detection unit 110 have mostly one high and narrow value as shown in the green portion of FIG. 3.

Also, the target of interest (red arrow in FIG. 2) is formed over a large area with a small frequency in the histogram (red box in FIG. 3). In other words, the infrared image obtained through the infrared detection unit 110 has a very low contrast ratio.

Therefore, it is necessary to improve the contrast ratio of the infrared image in the image signal processing unit 120. When improving the contrast ratio, it is preferable to convert a high dynamic range of 12 bits to 14 bits detected by the infrared detection unit 110 into 8 bits usable by the image output unit 130.

Histogram Stretching, Histogram Equalization, and Plateau Equalization are typically used as conventional techniques for improving the contrast ratio of the image signal processor 120.

The histogram stretching technique is also called AGC (Automatic Gain Control). This eliminates the extreme values of the histogram and linearly compresses the high dynamic range into the 8-bit domain. However, this method has a limitation of not being able to effectively express the detailed area of the thermal image.

The histogram smoothing technique calculates the mapping from the input cumulative distribution function (CDF) to the output cumulative distribution function. Histogram smoothing results in excessive contrast improvement when a high peak is present in the input histogram, resulting in a noisy, washed-out output image.

The flatto smoothing technique is a technique that reduces noise amplification in a flat region by removing the high peak value of the input histogram as a threshold. This method does not take into account the distribution of the input histogram. Therefore, the output image becomes unnatural because it uses the same threshold in the histogram.

The present invention aims to solve the above and other problems. Another object is good for viewing without artificiality in the process of converting an infrared or visible light image with a high dynamic range of 12bit to 14bit into 8bit, or in the process of improving the contrast of infrared light with low contrast or visible light. It is an object of the present invention to provide an image processing method and a gradient distribution based image processing method for performing a contrast ratio improvement method for calculating a natural output image in real time.

According to an aspect of the present invention to achieve the above or another object, an image input unit for receiving an image; An input histogram calculator calculating a histogram of the received image; A histogram sorting unit to sort the histogram using a sorting algorithm; An inclination distribution calculator for calculating an inclination distribution for the sorted histogram; A target histogram calculation unit calculating a target histogram through the calculated gradient distribution; And a contrast ratio improvement unit that rearranges the target histogram to calculate a transfer function, and outputs an image with an improved contrast ratio using the transfer function.

In an embodiment, the histogram sorting unit may sort the histogram calculated by the input histogram calculator in ascending order using the sorting algorithm.

In another embodiment, the gradient distribution calculator may calculate the gradient distribution based on the total number of horizontal pixels, vertical pixels, and unique pixel values of the input image.

In another embodiment, the target histogram calculator may calculate the target histogram by applying a weight between the aligned histogram and the slope distribution.

In another embodiment, the target histogram calculation unit may calculate a small value as the target histogram by comparing the aligned histogram and the slope distribution.

In another embodiment, the contrast enhancement unit may calculate the transfer function by rearranging through the mapping table calculated by the histogram sorting unit. At this time, the cumulative distribution function can be used as the transfer function.

In addition, according to another aspect of the present invention, receiving an image; Calculating a histogram for the input image; Sorting the histogram using a sorting algorithm; Calculating a slope distribution for the sorted histogram; Calculating a target histogram through the calculated gradient distribution; And re-sorting the target histogram to calculate a transfer function, and outputting an image with an improved contrast ratio by using the transfer function.

In an embodiment, sorting the histogram may include sorting the histogram calculated for the input image in ascending order using the sorting algorithm.

In another embodiment, the calculating of the gradient distribution may include: calculating the gradient distribution based on the total number of horizontal and vertical pixels, and unique pixel values of the received infrared image; It may include.

In another embodiment, calculating the target histogram may include calculating the target histogram by applying a weight between the aligned histogram and the gradient distribution.

In another embodiment, calculating the target histogram may include comparing the ordered histogram with the gradient distribution, and calculating a small value as the target histogram.

In another embodiment, the step of outputting the image with the improved contrast ratio; calculating the transfer function by re-sorting through the mapping table calculated in the histogram sorting step. At this time, the cumulative distribution function can be used as the transfer function.

When explaining the effects of the image processing method and the gradient distribution based image processing according to the present invention are as follows.

According to at least one of the embodiments of the present invention, when processing a contrast ratio of an input digital image signal in an infrared thermal imaging device or a visible light imaging device, there is an advantage that a high-quality display can be enabled by using a gradient distribution.

Further scope of applicability of the present invention will become apparent from the following detailed description. However, various changes and modifications within the spirit and scope of the present invention may be clearly understood by those skilled in the art, and thus, it should be understood that specific embodiments such as detailed description and preferred embodiments of the present invention are given as examples only.

1 is a block diagram illustrating an embodiment of a conventional infrared thermal imaging device.
2 is a view for explaining an embodiment of the infrared image generated by the conventional infrared thermal imaging device according to FIG.
3 is a graph showing the histogram of FIG. 2.
4 is a block diagram illustrating an embodiment of an image processing apparatus based on a gradient distribution according to the present invention.
5 is a flow chart for explaining an embodiment of a gradient distribution-based image processing method according to the present invention.
6 is a graph for explaining an embodiment of an aligned histogram and a target histogram.
7 is a graph for explaining an example of a histogram matching-based target histogram.
8 is a graph for explaining an embodiment of the transfer function.

Hereinafter, exemplary embodiments disclosed herein will be described in detail with reference to the accompanying drawings, but the same reference numerals will be assigned to the same or similar elements regardless of reference numerals, and redundant descriptions thereof will be omitted. The suffixes "modules" and "parts" for the components used in the following description are given or mixed only considering the ease of writing the specification, and do not have meanings or roles distinguished from each other in themselves. In addition, in describing the embodiments disclosed in this specification, detailed descriptions of related well-known technologies are omitted when it is determined that they may obscure the gist of the embodiments disclosed herein. In addition, the accompanying drawings are only for easy understanding of the embodiments disclosed in the present specification, and the technical spirit disclosed in the specification is not limited by the accompanying drawings, and all modifications included in the spirit and technical scope of the present invention , It should be understood to include equivalents or substitutes.

Terms including ordinal numbers such as first and second may be used to describe various components, but the components are not limited by the terms. The terms are used only for the purpose of distinguishing one component from other components.

When an element is said to be "connected" or "connected" to another component, it is understood that other components may be directly connected to or connected to the other component, but other components may exist in the middle. It should be. On the other hand, when a component is said to be "directly connected" or "directly connected" to another component, it should be understood that no other component exists in the middle.

Singular expressions include plural expressions unless the context clearly indicates otherwise.

In this application, the terms "comprises" or "have" are intended to indicate the presence of features, numbers, steps, actions, components, parts or combinations thereof described herein, one or more other features. It should be understood that the existence or addition possibilities of fields or numbers, steps, operations, components, parts or combinations thereof are not excluded in advance.

Hereinafter, the present invention will be described in more detail with reference to the accompanying drawings. Not all embodiments of the present invention are disclosed in the following description. The invention can be implemented in a wide variety of forms, and should not be construed as limited to the embodiments disclosed herein. These embodiments are provided to meet legal requirements for filing. The same reference numerals are used throughout the same components.

The inclined distribution-based image processing apparatus and image processing method according to the present invention can be applied to both infrared thermal imaging devices and visible light cameras. Hereinafter, as an example, an infrared thermal imager is mainly described, but is not limited thereto.

4 is a block diagram illustrating an embodiment of an image processing apparatus based on a gradient distribution according to the present invention.

4, the gradient distribution-based image processing apparatus according to the present invention includes an image input unit 410, an input histogram calculation unit 420, a histogram alignment unit 430, a gradient distribution calculation unit 440, and a target histogram calculation. It may include a portion 450 and the contrast ratio improving unit 460.

The image input unit 410 may receive, for example, an infrared image. For example, 12-bit to 14-bit infrared images may be input, but are not limited thereto. That is, an infrared image having various bits can be input. Also, an image by a general visible light camera having various bits may be input.

The input histogram calculation unit 420 calculates a histogram for the input image.

The histogram sorting unit 430 sorts the histogram using a sorting algorithm.

The gradient distribution calculator 440 calculates a gradient distribution for the sorted histogram.

The target histogram calculator 450 calculates the target histogram through the calculated gradient distribution.

The contrast ratio improvement unit 460 rearranges the target histogram to calculate a transfer function, and outputs an image with an improved contrast ratio using the transfer function.

In an embodiment, the histogram sorter 430 may sort the histogram calculated by the input histogram calculator 420 in ascending order using the sort algorithm.

In another embodiment, the gradient distribution calculation unit 440 may calculate the gradient distribution based on the total number of horizontal pixels, vertical pixels, and unique pixel values of the input image.

In another embodiment, the target histogram calculator 450 may calculate the target histogram by applying a weight between the aligned histogram and the slope distribution.

In another embodiment, the target histogram calculator 450 may compare the sorted histogram with the slope distribution to calculate a small value as the target histogram.

In another embodiment, the contrast improvement unit 460 may calculate the transfer function by re-sorting through the mapping table calculated by the histogram sorting unit 430.

5 is a flowchart for explaining an embodiment of an image processing method based on a gradient distribution according to the present invention.

6 is a graph for explaining an embodiment of an aligned histogram and a target histogram.

7 is a graph for explaining an example of a histogram matching based target histogram.

8 is a graph for explaining an embodiment of the transfer function.

Referring to FIG. 5, a step (S510) of receiving an infrared image is performed.

Specifically, an infrared image having a high dynamic range of 12 to 14 bits received from the infrared detection unit 110 may be input. This is an embodiment, and is not limited thereto. That is, an infrared image having various bits and an image obtained by a general visible light camera may be received.

 The image input unit 410 stores 12 to 14 bit infrared images output from the infrared detection unit 110 in a frame memory to enable image processing. As described above, it is not limited to 12 bits to 14 bits.

Subsequently, a step S520 of calculating an input histogram is performed.

The input histogram calculator 420 calculates a histogram distribution for the input digital image signal stored in each frame period in the frame memory. 3 shows a histogram distribution chart for image input of infrared thermal imaging equipment. Referring to FIG. 3, the input digital image signal of the infrared thermal imaging equipment is not evenly distributed over the entire band, but is characterized by being concentrated and distributed in a certain band.

Then, a step (S530) of sorting the histogram using the sorting algorithm is performed.

The histogram sorter 430 sorts the histogram calculated by the input histogram calculator 420 in ascending order using a sorting algorithm. 6 shows a histogram distribution chart arranged in ascending order. (Blue graph in FIG. 6, Sorted, Target) Here, when sorting in ascending order, the value whose histogram distribution is '0' is removed so that only unique pixel values (gray levels) are sorted.

Next, a step (S540) of calculating the slope distribution using the aligned histogram is performed.

The gradient distribution calculation unit 440 calculates the gradient distribution using the aligned histogram of the histogram alignment unit 430. Here, the slope distribution is a distribution that increases with a constant slope, and can be calculated as in Equation 1 below.

[Equation 1]

,

,

은 경사 분포이며, W는 영상의 가로 픽셀 수, H는 세로 픽셀 수, L은 고유한 픽셀 값의 총 개수이다. here,

Is a gradient distribution, W is the number of horizontal pixels in the image, H is the number of vertical pixels, and L is the total number of unique pixel values.

Thereafter, a step (S550) of calculating a target histogram through the calculated gradient distribution is performed.

The target histogram calculator 450 calculates the target histogram through the calculated gradient distribution. Various methods can be used to calculate the target histogram. In the present invention, a method for calculating a target histogram using an optimization technique and a method for calculating a target histogram using histogram matching will be described.

First, in the method of calculating the target histogram using the optimization technique, the target histogram is calculated through the weight of the slope distribution and the aligned histogram as shown in Equation 2 below.

[Equation 2]

는 목표 히스토그램이며,

는 정렬된 히스토그램(Hs)과 경사 분포 간의 가중치이다.

가 0 이면 목표 히스토그램은 정렬된 히스토그램 분포와 동일하며(도 6에서 파란색 그래프, Sorted),

가 커질수록 경사 분포를 따라가게 된다. (도 6에서 자주색과 하늘색 그래프,

가 1 또는 2일 때)here,

Is the target histogram,

Is the weight between the ordered histogram (Hs) and the slope distribution.

If is 0, the target histogram is the same as the sorted histogram distribution (blue graph in FIG. 6, Sorted),

As it increases, it follows the slope distribution. (Purple and sky blue graph in FIG. 6,

Is 1 or 2)

Second, in the method of calculating the target histogram using histogram matching, the target histogram is calculated using a small value through comparison between the aligned histogram and the inclined distribution.

에 가중치

를 반영하여 다양한 목표 히스토그램을 얻을 수도 있다. 가중치

가 커질수록 히스토그램 평활화와 동일한 영상을 만들어 낸다. (도 7에서 초록색 그래프, Target2)At this time, the slope of the slope distribution

Weight on

It is also possible to obtain various target histograms by reflecting. weight

As it increases, it creates the same image as the histogram smoothing. (Green graph in Figure 7, Target2)

Next, a step (S560) of outputting an image with an improved contrast ratio proceeds.

The contrast ratio improvement unit 460 recalculates the target histogram calculated by the target histogram calculation unit 450 through the mapping table calculated by the histogram alignment unit 430 to calculate a transfer function (refer to FIG. 8), and delivers Using the function, an image with improved contrast is output. At this time, the cumulative distribution function can be used as the transfer function.

When explaining the effects of the image processing method and the gradient distribution based image processing according to the present invention are as follows.

According to at least one of the embodiments of the present invention, when processing a contrast ratio of an input digital image signal in an infrared thermal imaging device or a visible light imaging device, there is an advantage that a high-quality display can be enabled by using a gradient distribution.

The above detailed description should not be construed as limiting in all respects, but should be considered illustrative. The scope of the present invention should be determined by rational interpretation of the appended claims, and all changes within the equivalent scope of the present invention are included in the scope of the present invention.

Claims (12)

An image input unit that receives an image;
An input histogram calculator calculating a histogram of the received image;
A histogram sorter that generates a mapping table for the histogram and sorts the histogram using a sorting algorithm;
An inclination distribution calculator for calculating an inclination distribution for the sorted histogram;
A target histogram calculator calculating a target histogram using the sorted histogram and the calculated slope distribution; And
And a contrast ratio improvement unit for rearranging the target histogram using the mapping table and outputting an image with an improved contrast ratio using a transfer function that converts the rearranged target histogram into an image. Based image processing device. According to claim 1,
The histogram alignment unit,
An image processing apparatus based on a gradient distribution, characterized in that the histogram calculated by the input histogram calculation unit is sorted in ascending order using the sorting algorithm. According to claim 1,
The gradient distribution calculation unit,
A gradient distribution-based image processing apparatus characterized in that the gradient distribution is calculated based on the number of horizontal pixels, vertical pixels, and the total number of unique pixel values of the input image. According to claim 1,
The target histogram calculation unit,
An image processing apparatus based on a gradient distribution, wherein the target histogram is calculated by applying a weight between the aligned histogram and the gradient distribution. According to claim 1,
The target histogram calculation unit,
Comparing the sorted histogram and the gradient distribution, a gradient distribution-based image processing apparatus characterized by calculating a small value as the target histogram. delete Receiving an image;
Calculating a histogram for the input image;
Generating a mapping table related to the histogram;
Sorting the histogram using a sorting algorithm;
Calculating a slope distribution for the sorted histogram;
Calculating a target histogram using the aligned histogram and the calculated slope distribution;
Rearranging the target histogram using the mapping table; And
And outputting an image with an improved contrast ratio by using the transfer function that converts the re-aligned target histogram into an image. The method of claim 7,
Sorting the histogram;
And sorting the histogram calculated for the received image in ascending order using the sorting algorithm. The method of claim 7,
Computing the gradient distribution;
And calculating the gradient distribution based on the number of horizontal pixels, vertical pixels, and the total number of unique pixel values of the received image. The method of claim 7,
Computing the target histogram;
And calculating the target histogram by applying a weight between the aligned histogram and the gradient distribution. The method of claim 7,
Computing the target histogram;
And comparing the ordered histogram and the gradient distribution to calculate a small value as the target histogram. delete

Images