Study On Small Target Detection Algorithms In Infrared Images

Infrared imaging missile guiding has been extensively used in guided missile due to its strong all-weather performance, high guiding precision and strong resistance to disturbances, it is also one of the leading techniques in missile guiding currently. Targets far away often appear as small target in infrared images with a tiny area, weak energy, blurred boundary, and even the deficiency of texture information. Therefore to detect small targets precisely in an infrared image has been a key technique in infrared missile guiding. This dissertation studies on small target detection algorithms in infrared images. Firstly it sums up the current infrared small target detection techniques and gives brief introductions to some major algorithms. Based on the previous work, the major contribution of this paper are like the following:1. Gives an intensive analysis of small target detection algorithm using TDLMS filter and points out the defects cause by the limitations of the filter. A new algorithm called "small target detection using TDLMS based on neighborhood analysis" is proposed, which improves the performance of the small target detection scheme using TDLMS filter.2. Studies on a small target detection algorithm known as "Eigentarget" using PCA. Based on this method, new algorithms using PPCA and PKPCA are proposed. Together they forms detection algorithm based on PCA family. Experimental result analysis and performance comparisons are given.3. For the special application in infrared linear detector imaging, a new target detection idea based on column detection is proposed, which improves the efficiency for small target detection in infrared linear detector imaging. Three different detection algorithms are studied under the proposed idea. Experimental result analysis and performance comparisons are given.