Research And Implementation Of Infrared Image Adaptive Enhancement Algorithm In Dynamic Scene

When infrared search and tracking equipment and thermal imager collect video images,dynamic infrared video images will constantly switch between complex scenes and even scenes in real time.At this time,the algorithm that can be used to enhance the details of infrared image is not applicable in this case.Therefore,it is very important to study an infrared image enhancement algorithm which can automatically adapt to a variety of scenes.Based on the infrared video image captured by infrared thermal imager,this paper studies the adaptive enhancement algorithm of infrared video image in dynamic scene.Firstly,the initial input image is smoothed by the guide filter,and the difference between the initial input image and the smoothed image is made to obtain the based layer image containing large dynamic temperature information and the detailed layer image containing small dynamic temperature information;secondly,the base layer image is processed by adaptive dynamic compression and the detail layer image is controlled by adaptive gain;finally,the adaptive is determined by the scene The fusion coefficient is used to fuse the base layer image and the detail layer image to obtain the output image.In order to reduce the dynamic compression time of the image in the base layer,highlight the details of the image,reduce the impact of the noise in the details layer on the output image and achieve the effect of adaptive scene,a one-dimensional compression array is designed to compress the image in the base layer dynamically by using the adaptive threshold parameters and histogram distribution information that can filter the effective gray values;according to the weight linear coefficient in the guide filter The characteristics of the filter window(when there is an obvious edge in the filter window,the weight coefficient approaches to 1,when there is a smooth area in the filter window,the weight coefficient approaches to 0).The linear coefficient of the weight determined when guiding the filter to smooth the initial input image is introduced into the processing of the detail layer image as the adaptive control coefficient of the detail layer image,so as to suppress noise and retain the edge Objective: to obtain the number of effective gray values in the histogram according to the determined threshold value,to roughly judge the scene information of the image by the ratio between the number of effective gray values and the total gray values,and to determine the self-adaptive fusion proportion coefficient of the basic layer image and the detail layer image according to the different scene information,so as to realize the image fusion of the adaptive scene.The experimental results are compared with the image processed by histogram equalization algorithm,infrared image detail enhancement algorithm based on bilateral filtering and adaptive infrared image enhancement algorithm based on guided filtering.Nine different scenes are selected to analyze from subjective and objective levels.The results show that the image processed by this algorithm can highlight the detail contour information and reduce the influence of detail layer noise on the fused output image.Compared with the other three algorithms,the processed image can give a better visual effect.According to the objective evaluation,the processing time of the algorithm is 1.882 seconds,which can make the video infrared image play smoothly.