In order to improve the image quality, infrared images are often digitally processed by computer. Factors that usually affect the quality of infrared images include fixed noise interference, random noise interference, and differences in responsivity. When the computer is used to process the image, the noise can be suppressed and the inhomogeneity can be compensated, thereby improving the image quality and the accuracy of temperature observation. Since the random noise of the image is additive noise, the frames are uncorrelated and the mean value is zero, the multi-frame averaging method can improve the signal-to-noise ratio of the image. Differences in responsivity can be compensated for by pixel-by-pixel responsivity correction of the image input to the computer. For the suppression of fixed noise, the method of inter-frame subtraction can be used to eliminate it, and it is very convenient for the computer to complete the image subtraction operation.

 

Image enhancement methods can be divided into three categories: temporal domain processing, spatial domain processing and transform domain processing. Time domain enhancement includes time delay integration, inter-frame comparison and other methods; spatial domain enhancement is divided into point processing and neighborhood processing. The former includes contrast stretching, histogram processing and other methods, and the latter commonly used median filtering, mean value Transform domain enhancement is to perform various filtering on the basis of image transformations such as discrete Fourier transform and wavelet transform, and finally achieve the purpose of enhancement.

 

The pros and cons of an image enhancement algorithm are not absolute. Due to the different purposes and requirements of specific applications, the required augmentation techniques are also quite different. Therefore, fundamentally, there is no general standard for image enhancement, and the observer is the final judge of the quality of a certain enhancement method. The processing effect of the enhancement algorithm is not only related to the algorithm itself, but also directly related to the numerical characteristics of the image. In practical applications, an appropriate method should be selected according to the characteristics of image data and processing requirements.