A Study Of Technology For Complex Background Suppression And Detection Of The Weak And Small Target

In the shooting range test, due to the surrounding environment is changing, changes in the sky clouds, trees shelter, inadequate lighting, and other factors make the background becomes complex, and a smaller number of pixels imaged small-caliber projectile, only few pixels, the SNR is low, so it is difficult to detect the projectile. Therefore, how to detect the weak and small target under the complex background in practical application is an extremely meaningful topic.This paper describes the algorithm of detecting the weak and small target under the complex background. Analyzing the characteristics of the image, such as the background, the target and the noise, and establishing the mathematical model according to the image, and researching how to suppress the complex background based on the mathematical model. In order to suppress the complex background, using some different methods to achieve it, such as the weighted mean filtering, the median filtering, the spatial high-pass filtering and the improved morphological filtering. Among them, the method of the improved morphological filtering is better, and the SNR is improved. In the single image, using the algorithm to detect the weak and small target based on wavelet transform. In order to highlight details of the image, selecting the appropriate wavelet is first, and then analyzing the wavelet transform coefficients of each band after the transforming, and using the method of thresholding to detect the target. The algorithm is simple, and steps are reasonable, but it can not accurately detect the weak and small target for the single image, and it needs to be improved. Lastly, for the deficiencies of the single image, this paper uses a algorithm of the sequence images to detect the target. First, the wavelet transform is used to process the image, and then identifying the suspicious targets by threshold segmentation. Finally, the pipeline filtering in image sequences further conform the true weak and small target.The algorithm of the target detection in this paper provides a more effective way for all types of flying projectile image processing, and the method had been applied in the actual detected, and the performance of the algorithm is superior to the single-frame target detection, it can accurately detect the weak and small target under the complex background, and the effect is expected, it lay a good foundation for the extraction and testing of projectile density.