Fuzzy Infrared Image Quality Assessment Method

With the progress and development of society, image and video has become anindispensable part of modern life, so we want to obtain high-quality image and video.However, the image degradation problem may occur, because it may be affected bynoise in image capture, compression, processing, transmission and display process.Degraded image will have a serious impact on our visual perception. Image is animportant means when human perceptive information, express information or transferinformation. However, image quality has a direct impact on accuracy and reliability ofacquiring information. The purpose of image quality assessment is to measure thedegree of image degradation and reduce the difficulty of image later processing. Imagequality assessment can provide monitoring means in infrared image capture,processing, transport and other sectors, which has guiding significance for qualitativeevaluation of infrared imaging device performance.Firstly, the work principle of infrared thermal imaging and imaging infrared imagecharacteristic is discussed, and the analysis of advantages and disadvantages ofinformation entropy and structural similarity image quality assessment algorithm. Dueto the poor sensitivity calculation of information entropy, which can only statisticaloverall entropy of image, the image quality assessment algorithm based on fuzzyentropy is designed. This method considers of the important property of the uncertaintyof the image is often fuzzy, but not random. So the concept of fuzzy entropy isintroduced into the quality assessment of infrared image. The method has manyadvantages: the simple principle, the easy way to understand and the use of fuzzyuncertainty which human visual can capture. The simulation results show that themethod is an accurate and reliable infrared image sharpness quality assessmentalgorithm.Secondly, the global fuzzy entropy does not consider the pixel distribution in space,it also can’t measure the degree of blurring of the local area in the image. Therefore，the local fuzzy entropy-based no-reference blur infrared image quality assessment method is proposed. The simulation results show that the general performance of thismethod is better than MSE and PSNR.The result of image quality assessment do not match with the subjective evaluationwhen structural similarity algorithm evaluate blur image, especially the image are lessclear. The evaluation result is Unreliable. The gradient information of image can reflectimage edge and texture information. therefore, it is introduced to structural similarityassessment algorithm to evaluate the quality of image. The evaluation method based onstructural similarity image sharpness of gradient is designed. The simulation resultsshow that the method can accurately reflect the image sharpness and the assessmentresult is consistent with subjective quality evaluation method.