Research On Infrared Small Target Detection Under Sea-sky Background

The small target detection is an important content involved in long-distance target tracking under the complex sea-sky background. This paper mainly researches the methods of the dim and small target detection in infrared image with sea-sky background, The main contributions are given below:(1)The characteristics of the actual infrared image of small target for analysis. Depending on infrared imaging sensors work environment, the background image classification. and a variety of different backgrounds(sky background, sea background and sea-sky background), real infrared images were analyzed and compared. Using different methods for the inhibition of different backgrounds, to improve the image background suppression capability, provide a theoretical basis.(2) Based on the infrared image' features, sea background, weak target, sky background characteristics are discussed in detail. Using the difference between the sky background and the sea background and the gray of the has the great different features distinguishing, this paper put forward to the information of the boundary of sky background and sea background by using the Hough Transform, which segments sky and sea. Using image texture feature, a time stack image is generated from the image and processed by Radon transform to find the forward wave direction in specific orientation, and wave direction to achieve effectively.(3)The paper focuses on the research and analysis the infrared image of gray level has a strong correlation between adjacent rows. After the method of line average subtracting are proposed, the nonlinear influence in temperature field could be inhibited. Inhibiting background and retaining high area of brightness are implemented by using morphological filtering. Judging out the remaining large area highlighted area for the waves of the sea, according to the direction of waves. Using skeleton extraction technology to eliminate and weak target size or larger wave edge. The experimental results show that the proposed algorithm greatly reduce unnecessary noise and interference waves.(4)The information of segmentation theoretical are analyzed and performance are compared. Two-dimensional maximum entropy have been presented combining with the threshold segmentation and the characteristic of the gray-scale distribution concentrated. The experimental results indicate that the algorithm can decrease the time consumption effectively, have good real-time capability and easy to engineering. It basically solved the problem of segmentation in the complex sea-sky background.In this paper, the infrared image of small target detection related technologies indepth research and analysis, and the use of sign infrared image of the proposed algorithm verified by experiments, experimental results show that the proposed algorithm can effectively improve the detection effect.