Infrared image contrast gain adjustment method, storage medium and infrared detector

Abstract

The invention discloses an infrared image contrast gain adjustment method, a storage medium and an infrared detector. The method comprises: 1) segmenting an input infrared image; 2) designing a fuzzy controller and determining the input variables and output variables of the fuzzy controller; 3) scaling the input variables and output variables of the fuzzy controller; 4) fuzzy processing the input variables; 5) performing fuzzy reasoning; 6) performing defuzzification; 7) calculating the gray value after contrast gain adjustment. According to the established fuzzy rule base, the infrared image contrast gain can be adjusted. The method has strong robustness, can effectively suppress infrared image noise, improve image contrast, and enhance infrared image detail information.

Description

Technical Field

[0001] The invention relates to an infrared image contrast gain adjustment method, a storage medium and an infrared detector. The method is based on a fuzzy control algorithm and realizes adjustment of the infrared image contrast gain according to an established fuzzy rule base. Background Art

[0002] With the rapid development of infrared technology, infrared imaging systems have been widely used in many fields. Benefiting from microelectronics technology, infrared imaging systems are developing in the direction of low power consumption, small size and multi-function. The current common infrared processing system consists of infrared detectors, FPGAs and other necessary components, which can obtain high-quality infrared images.

[0003] Infrared thermal imaging technology is to form an image by detecting the thermal radiation on the surface of an object. However, due to atmospheric attenuation, the infrared radiation energy reaching the optical lens is very small, the infrared image contrast is low, and the vision is relatively blurred. At the same time, when receiving the infrared radiation of the detected object, a large number of interference factors will also be received, such as ambient temperature, dust, atmospheric radiation, etc., resulting in large non-uniformity of the infrared image, high noise, unclear detail texture, low contrast, etc., which makes it difficult to meet people's requirements for high-quality infrared images, and the development of infrared technology is subject to certain restrictions.

[0004] At present, the technology and process of infrared focal plane detectors are not perfect enough. For example, factors such as random interference from the external environment, thermal balance of the environment, long transmission distances, and severe atmospheric attenuation will directly lead to problems such as low contrast and blurred image effects in infrared images. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide an infrared image contrast gain adjustment method, a storage medium and an infrared detector to address the shortcomings of infrared images, such as low contrast, blurred visual effects, large non-uniformity, high noise, and unclear detail texture.

[0006] The infrared image contrast gain adjustment method of the present invention is based on a fuzzy control algorithm. According to the established fuzzy rule base, the infrared image contrast gain can be adjusted. The method has strong robustness, can effectively suppress infrared image noise, improve image contrast, and enhance infrared image detail information.

[0007] The technical solution of the present invention to solve the above technical problems is:

[0008] A method for adjusting infrared image contrast gain comprises the following steps:

[0009] (1) Segment the input infrared image to obtain different dimensional information of the infrared image.

[0010] (2) Design a fuzzy controller and determine its input and output variables. The input variables of the fuzzy controller are parameters that can reflect the overall fluctuation range of the infrared image and parameters that can reflect the local information amount, so that the fuzzy controller can adjust the contrast gain of the infrared image.

[0011] (3) The input variables and output variables of the fuzzy controller are scaled and transformed into their respective domains. The scale transformation formula is as follows:

[0012]

[0013] (4) Perform fuzzy processing on the input variables to transform the precise values ​​of the input variables into fuzzy values.

[0014] (5) Perform fuzzy reasoning, divide both the input variables and the output variables into A fuzzy subsets, and establish a fuzzy rule base based on expert experience.

[0015] (6) Defuzzification: After fuzzy reasoning, the output variable obtained is still a fuzzy value. Defuzzification of the output variable converts it into an accurate output quantity, which can reflect the weight of the overall fluctuation of the infrared image and the weight of the local information amount of the infrared image.

[0016] (7) Calculate the grayscale value after contrast gain adjustment, and substitute the output into the following formula to obtain the grayscale value I after contrast gain adjustment:

[0017] I=(I ij -I mean )*G1+(I ij -I mean )*G2+I mean

[0018] Where: I ij is the gray value of the point, I mean is the mean gray value of the area, G1 is the difference weight, and G2 is the variance weight.

[0019] A computer-readable storage medium stores a computer program, which is executed by a processor to implement the steps of the infrared image contrast gain adjustment method.

[0020] An infrared detector containing the above-mentioned computer-readable storage medium.

[0021] The invention provides an infrared image contrast gain adjustment method, which effectively suppresses noise, improves the infrared image contrast, and enhances the infrared image detail information. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is a functional block diagram of Example 1.

[0023] Figure 2 Is the membership function of the input and output variables. In the figure: the horizontal axis is the domain of the input and output variables, the vertical axis is the membership degree, NB, NS, Z, PS, and PB are the five fuzzy subsets of the input and output variables.

[0024] Figure 3 This is a comparison diagram of the infrared image contrast gain before and after adjustment in Example 1, where (a) is the original image and (b) is the image after contrast gain adjustment. DETAILED DESCRIPTION

[0025] In order to solve the defects of low contrast, low clarity, poor ability to distinguish small target shapes and details of infrared images, the present invention proposes an infrared image contrast gain adjustment method. The infrared image contrast gain is adjusted by fuzzy control. The algorithm can adjust the infrared image contrast gain according to the established fuzzy rule base, so that it can effectively suppress infrared image noise, enhance infrared image details, and improve infrared image contrast.

[0026] In order to further illustrate the technical solution of the present invention, Example 1 is described in detail below.

[0027] Example 1

[0028] like Figure 1 As shown, this embodiment includes the following steps:

[0029] 1) Taking the division of an infrared image into several blocks as an example, the input infrared image of 768 rows and 1024 columns is divided into 64 infrared images of 12 rows and 16 columns.

[0030] 2) Take the design of two single-input and single-output type I fuzzy controllers as an example to enable them to realize the function of adjusting the contrast gain of infrared images. Since the gray value difference and variance can well reflect the contrast, detail information and noise information of the area, the input of the fuzzy controller is selected as the infrared image pixel difference D i and variance D var , the outputs are respectively the infrared image difference D i The weight G1 and the variance weight G2.

[0031] Assume that the gray value of each pixel is I ij , calculate the difference D between the maximum and minimum grayscale values ​​of each pixel of the cut infrared image according to the following formula: i and variance D var Among them, I max ,I minis the maximum and minimum grayscale value of the area, I mean is the mean gray value of the area.

[0032]

[0033] 3) Scale the input and output of the fuzzy controller to their respective domains. Quantization factor k a The calculation is as follows:

[0034]

[0035] Where: DD is the domain range, R I The value range.

[0036] 4) Then the input is fuzzy processed to convert the precise value of the input into a fuzzy value. A is a fuzzy set on R, μ A (x) is called the degree to which x belongs to the fuzzy set A, that is, the membership function of A. Since the Gaussian membership function has good smoothness, symmetry, and no zero point, the Gaussian membership function is taken as an example, and the input and output membership functions of the designed fuzzy controller are all Gaussian membership functions f(x).

[0037]

[0038] In the formula: σ represents the standard deviation, c represents the mathematical expectation value obeyed by the random variable X.

[0039] 5) Taking the input and output variables as examples, we divide them into five fuzzy subsets {PB, PS, Z, NS, NB}, and establish a fuzzy rule base based on expert experience. The difference between the maximum and minimum grayscale values ​​and the variance of the region can reflect the detailed information of the region, but the dimension of the variance is higher than that of the difference, and the information is more comprehensive. According to experience, when the difference is small, the noise in the region is strong, and the contrast gain needs to be suppressed to achieve the effect of suppressing noise; when the difference is neither large nor small, its detailed information needs to be enhanced, so the contrast gain needs to be increased; when the difference is large, the detailed information in the region is rich, and the contrast gain needs to be reduced, otherwise the noise will be amplified. Similarly, when the variance is small, the contrast gain needs to be reduced to suppress the noise; when the variance is neither large nor small, the contrast gain needs to be increased to enhance the detailed information of the region; when the variance is large, the detailed information in the region is rich, and there is basically no need to adjust the contrast gain, otherwise the noise may be amplified. Based on the above experience, the fuzzy rule base is established as shown in the following table. Taking the Mamdani If-Then rule as an example, the designed fuzzy controllers all use the MamdaniIf-Then rule.

[0040] G1 fuzzy rule table

[0041] <![CDATA[D i ]]> NB NS Z PS PB <![CDATA[G1]]> NB NS PS Z NS

[0042] G2 fuzzy rule table

[0043] <![CDATA[D var ]]> NB NS Z PS PB <![CDATA[G2]]> NB NS PB Z Z

[0044] In the above table, D i is the difference between the maximum and minimum grayscale values ​​of the infrared image pixels, D var is the variance, G1 is the difference D i The weight of G2 is the variance weight.

[0045] The membership functions of the input and output variables are as follows: Figure 2 As shown, the horizontal axis is the domain of input and output variables, the vertical axis is the membership degree, and NB, NS, Z, PS, and PB are the five fuzzy subsets of the input and output variables.

[0046] 6) After the input variables are fuzzified, the fuzzy rules that are activated can be deduced based on the established fuzzy rule base. After the activated fuzzy rules are subjected to fuzzy reasoning, the output value obtained is still a fuzzy value, and a defuzzification step is required to convert it into an accurate output quantity. Taking the centroid method for defuzzification as an example, the defuzzification formula is as follows.

[0047]

[0048] Where: u N (x i ) is x i The membership degree, x i is the horizontal coordinate corresponding to the center of the area.

[0049] 7) After deblurring, the exact values ​​of the block difference weight G1 and variance weight G2 of the pixel point ij are obtained. Substituting them into the following formula, the grayscale value I after contrast gain adjustment can be obtained.

[0050] I=(I ij -I mean )*G1+(I ij -I mean )*G2+I mean

[0051] Where: I ij is the gray value of the point, I mean is the mean gray value of the area, G1 is the difference weight, and G2 is the variance weight.

[0052] Depend on Figure 3It can be seen that the infrared image contrast gain adjustment method based on fuzzy control can effectively enhance the contrast and detail information of the infrared image, and has a good suppression effect on the noise of the infrared image, effectively improving the quality of the infrared image.

Claims

1. A method for adjusting contrast gain of an infrared image, characterized in that: The following steps are involved: (1) Divide the input infrared image into several blocks to obtain information of different dimensions of the infrared image; (2) Design two single-input single-output fuzzy controllers and determine the input variables and output variables of the fuzzy controllers. The input variables of the fuzzy controllers are parameters that can reflect the overall fluctuation range of the infrared image and parameters that can reflect the local information volume, respectively, so that the fuzzy controllers can realize the function of adjusting the contrast gain of the infrared image; (3) The input variables and output variables of the fuzzy controller are scaled and transformed into their respective domains. The scale transformation formula is as follows: Where: k a is the quantitative factor, DD is the domain range, R I is the value range; (4) Perform fuzzy processing on input variables to transform the precise values ​​of input variables into fuzzy values; (5) According to the Mamdani fuzzy reasoning method, fuzzy reasoning is performed, and both the input variables and the output variables are divided into A fuzzy subsets. The If-Then rules are established as follows: G1 fuzzy rule table G2 fuzzy rule table In the above table, D i D is the difference between the maximum and minimum grayscale values ​​of the infrared image pixels in each block after segmentation. var is the gray value variance of the infrared image pixel point in each block after segmentation, G1 is the difference D i The weight of G2 is the variance D var The weight of (6) Defuzzification: After fuzzy reasoning, the output variable obtained is still a fuzzy value. Defuzzification of the output variable converts it into an accurate output quantity, which can reflect the weight of the overall fluctuation of the infrared image and the weight of the local information amount of the infrared image; (7) Calculate the grayscale value after contrast gain adjustment, and substitute the output into the following formula to obtain the grayscale value I after contrast gain adjustment: I=(I ij -I mean )*G1+(I ij -I mean )*G2+I mean Where: I ij is the original gray value of pixel ij, I mean is the mean gray value of the block where the pixel point ij is located after segmentation.

2. The infrared image contrast gain adjustment method according to claim 1, characterized in that: The method for fuzzifying the input variables in step (4) includes any one of the following: the graded fuzzy set method, the input point membership 1 method, the single point fuzzy set method and the membership method.

3. The infrared image contrast gain adjustment method according to any one of claims 1 to 2, characterized in that: The defuzzification algorithm in step (6) includes any one of the maximum membership method, the centroid method and the median method.

4. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer program is executed by a processor to implement the steps of the infrared image contrast gain adjustment method as described in any one of claims 1 to 3.

5. An infrared detector comprising a computer-readable storage medium as claimed in claim 4.

Images