The Research On Dual-band Infrared Target Detection Algorithm

With the increasing demand for indicators such as the distance of infrared detection systems and the increasingly complex operational environment,the difficulty of detecting and identifying infrared targets has also increased.It is difficult to accurately detect and identify the target using only the image information obtained by a single infrared detector.The infrared target detection algorithm based on dual-band image fusion utilizes the redundancy and complementarity of image information in different bands.Through information fusion,the probability of infrared target detection and recognition is improved,and the objectivity and correctness of detection and recognition are further improved,which has become a research hotspot in the field of infrared detection.First of all,based on the infrared image feature description model,the image analysis is carried out from three aspects,including the overall brightness of the image,the richness of the detail information and the target background contrast.Using specific infrared images,the mean and standard deviation indicators were used to analyze the imaging characteristics of medium and long-wave infrared targets and small targets,respectively.Secondly,a dual-band infrared target fusion detection algorithm based on NSCT-PCNN and Canny edge detection is proposed.The algorithm achieves pixel-level fusion of dual-band surface target images by combining NSCT with improved PCNN,enhancing the image details of the surface target and improving the overall clarity of the image.Then the Canny operator is used to carry out edge detection on the fusion result to extract the image contours,and the target and background are distinguished according to the closed characteristic of the contour curve,so as to locate the target physical position and achieve the purpose of surface target detection.For the difficulty of detection of weak and small targets,this paper studies the basic strategy of detecting re-fusion first.In the detection stage,a background suppression algorithm based on improved Singular Value Decomposition and Fixed Weighted Filtering is proposed.The algorithm enhances the small target by performing nonlinear transformation on the singular value matrix of the image decomposition,and implements suppression of complex background through fixed weighted filtering.Then the position and number of candidate targets are obtained by adaptive threshold segmentation.In the fusion phase,a dual-band weak and small target fusion algorithm based on Boolean logic and local gray discrimination is proposed.The algorithm realizes the screening of the real target by detecting the neighborhood gray distribution of the candidate target's center point.The proposed algorithm is simulated and analyzed by the acquired dual-band weak target image.The experimental results show that the proposed background suppression algorithm can effectively improve the efficiency of low-intelligence and low-noise detection of infrared images under low signal-to-noise ratio,and the fusion algorithm can further improve the target detection performance of the system.