Research On Target Detection And Location Technology Based On Image Fusion

The detection and location technology of moving object in video image sequence has been applied more and more.However,the accuracy and robustness of the existing target detection and location algorithms are still affected by the complex background in the image,such as the change of illumination in the target area,the small contrast between the target and the background,and the serious occlusion of the target.For this reason,this paper proposes a target detection and location technology based on image fusion for the influence of illumination changes in the target area.In order to eliminate the influence of illumination change in the image,an image enhancement method based on bilateral filtering is designed in this paper.The illumination image of the source image is obtained by bilateral filtering,and then the reflected image is obtained according to Retinex theory.Different processing strategies are adopted to enhance the illumination image and the reflected image.Compared with two classical enhancement algorithms from subjective and objective aspects,the information entropy,mean and standard deviation of the proposed method are better than the other two methods.To eliminate the effects of illumination changes,a high dynamic range image fusion method based on different gain images is designed.Firstly,the fusion coefficient is determined by testing the electronic parameters of the camera,and then the high dynamic range image is obtained.After gray stretching,the displayable image is obtained.Secondly,this method is compared with two traditional fusion algorithms from subjective and objective aspects,at the same time,the dynamic range of the fusion processed camera is compared with the general camera.Chenxin G2020 Complementary Metal Oxide Semiconductor(CMOS)Image Sensor is selected to complete the experimental test.The dynamic range of the camera in high-gain mode and low-gain mode before fusion is 59.2dB and 66.1dB respectively.The dynamic range of the camera can reach 87.5dB after processing by this method,and the imaging quality is good.The method proposed in this paper has less computation and is easy to implement in hardware.In order to improve the accuracy of the algorithm,this paper combines the improved centroid tracking algorithm with the mean shift algorithm to reduce the error between the center position of the tracking target and the real center position of the target.Experiments show that the performance of the improved algorithm is increased.Finally,an experimental platform for image acquisition is built to verify the overall algorithm.The experimental results show that the average similarity coefficient of the improved target tracking algorithm based on fusion image is about 30% higher than that based on high-gain image,and about 60% higher than that based on low-gain image.