Study on Target Detection in Infrared Images Based on Vision Saliency

摘要

Automatic target detection of images is an important key point for the application of infrared images. By utilizing the contrast and saliency of computer vision, a new method is proposed to detect targets in infrared images. Objects are extracted firstly by Classical Fuzzy C-means cluster and connection analysis. Then target objects and background objects are distinguished by computing vision saliency. The parameters object-based analysis are derived and carried out for each object, and then these objects are filtered by defining criteria conditions. The results show that the proposed algorithm is an effective method to detect targets in infrared images.