Target Detection And Location Technology In Complex Background

This paper mainly focuses on the target detection and location technique in the complex background infrared images.Based on the study of the methods of target detection and location both at home and abroad, and aimed at the complex background infrared images of our project, some effective detection methods have been proposed. Among of the methods, the first improved algorithm based on background prediction is performed for small target detection in the complex background of infrared images. To begin with, divide the pixels of the surrounding background into four areas with the pixel as the center, then use the part pixels of the four areas to predict the current pixel, next, take the maximum as the final predicted value, finally, remove the background and detect target. The second one is a method using the theory of mathematical morphology to detect the biggish target. In the method, first select the best structuring element; then, carry through background prediction and get the background image; next, subtract it from the original image; finally use the segmentation of setting threshold to complete image self-segmentation and get the target. The third one aims at the infrared serial images, though the Sequential Similarity Detection Algorithms (SSDA) to get the pixel coordinate of target. These detection results show that the three methods much more effective, high accuracy and good characteristic for real time.Detection process is simulated with the language matlab and C++ Builder and the results have shown perfect. Advantages and disadvantages of each algorithm are discussed at the end of the paper and the improved advice is also provided.