Research On Infrared Image Segmentation Based On Active Contour Model

Infrared imaging technology has been widely applied in military and civil fields, while the infrared image segmentation aims to extract interesting parts by utilizing relevant characteristics, and the segmentation result is of great importance to the process of target recognition and tracking. In recent years, active contour models(ACM) have been widely used in image segmentation, and some researchers have also tried to apply them to infrared image segmentation, but generally speaking, ACMs for infrared image segmentation are still in the exploration stage. On that account, this paper carries out corresponding research on the application of ACMs in the infrared image segmentation.After studying the existing results of the application of ACMs in image segmentation, aimed at the blurred boundaries, low intensity contrast and high noise of infrared images, a modified ACM driven by local entropy and local standard deviation is proposed in this paper, by taking advantage of the filter function of the standard deviation and the characteristic that gray-scale distribution can better measure the contrast. To establish the model more accurately, a new signed pressure force (SPF) function is constructed, we firstly construct the feature vector with the local entropy and local standard deviation corresponding to each pixel in the image, then we determine the evolution direction and magnitude of driving force for each pixel on the contour with the feature vectors, which can drive the evolution more accurately.Finally, we take advantage of the modified ACM to conduct experiments on the images taken by the infrared camera of the laboratory, and the simulation results demonstrate that the modified model can obtain more accurate and complete results for infrared images segmentation.