| Nowadays,the development of military science and technology has greatly affected the pattern and direction of war.The Infrared small target detection technology palys a vital role in the fields of infrared remote warning,infrared imaging guidance and infrared search tracking.The infrared imaging system is affected by the weather,clouds,selection of view.The interference of clutter in the infrared image is more,which makes the detection more difficult.Thus,the infrared small target detection technology in complex background is regarded as a difficult,challenging and significant scientific research topic.With the continuous innovation of infrared small target detection technology the infrared small target detection system itself has more stringent requirements on its effectiveness and power consumption.In this paper,infrared small target detection algorithm is studied in the field of infrared small target gray information,small target edge information and information entropy.This paper chooses Zynq heterogeneous platform as the processor of infrared small target detection system.The main contents include: 1.Analyze the architecture and design advantages of the Zynq platform,and design a small infrared target detection system based on Zynq.2.Study the infrared small target image on three aspects: gray distribution,small target edge information and image entropy.Analysis the expression of small targets in images,highlight the infrared small target feature structure,and detect the infrared small target from the feature.3.A new real-time algorithm of small infrared target detection for weighted local entropy and multi-scale morphological image fusion is proposed.The algorithm improves the performance of the single feature infrared small target detection algorithm through the idea of multi-feature fusion.Starting from the small target gray distribution characteristics,firstly,the multi-scale improved morphological operation is calculated,and the new differential compensation idea is combined to achieve the purpose of suppressing the complex background and enhancing the saliency of the infrared small target.At the same time,from the image entropy characteristic,the image entropy of the infrared small target is extracted to enhance the small target saliency.Then,the multi-feature fusion idea is used to enhance the infrared small target from different features.At the same time,the infrared small targets in different backgrounds are simulated and compared with other algorithms studied to verify the performance advantages of the proposed algorithms.4.The proposed infrared small target detection algorithm is transplanted into the infrared small target detection system based on Zynq to realize the infrared small target detection task.At the same time,combined with Vivado HLS,speed up the development of algorithms and realize the hardware acceleration of algorithms.The IP core of the new algorithm proposed in this paper is designed and implemented.And the IP core is simulated,tested,and verified by the hardware platform.The experimental results prove the effectiveness of the software and hardware design of the detection system. |
