We apply graph matching method to detect infrared small moving targets using image sequences. Candidates (interest points) detected in the first frame form one graph and the same candidates in the last frame form another one. The real moving targets are extracted by matching these two graphs. Experimental results demonstrate that the proposed method is robust and efficient to the translation and rotation of the background.

We present a particle filter (PF)-based algorithm to detect and track maneuvering infrared weak multiple targets at different signal-to-noise ratios for the scenes with the multiple targets number unknown and varying. A detecting filter and a tracking filter based on sequential likelihood ratio (LR) testing with fixed sample size are designed, respectively, for capturing new target and tracking confirmed targets. The algorithm is optimized with selectively particles sampling and adaptive process noise. Targets birth and death time are accurately estimated according to the change degree of the LR along with the corresponding state amended through PF backward recursion. Simulation results show that it is positive to detect and track maneuvering infrared weak multiple targets with the appearance and disappearance of more than one, which also achieves a significant improvement in state estimation especially for the time targets which appear and disappear.

This letter presents a correlation tracking algorithm based on edge detection, in allusion to the phenomenon that the target is susceptible to interference in long-wave infrared image. Firstly, the collected imageis processed by median filtering to remove the defects in the detector. Then thegradient information of the background is filtered out by edge detection method. Secondly, to enhance the edge information, the image is dealt with mathematical morphology (MM). And finally, the correlation matching method is done to acquire the miss-distance information of the target in the image. Through theoretical simulation and practical verification, it is proved that the algorithm has a better effect on inhibiting most of the background information, protecting the target's information effectively, enlarging the weight of the target's information in the correlation operation and improving stability of the detector.

The mean-shift algorithm has achieved considerable success in object tracking due to its simplicity and efficiency. Color histogram is a common feature in the description of an object. However, the kernel-based color histogram may not have the ability to discriminate the object from clutter background. To boost the discriminating ability of the feature, based on background contrasting, this letter presents an improved Bhattacharyya similarity metric for mean-shift tracking. Experiments show that the proposed tracker is more robust in relation to background clutter.

Synthetic aperture integral imaging provides the ability to reconstruct partially occluded objects from multi-view images. However, the reconstructed images suffer from degraded contrast due to the superimposition of foreground defocus blur. We propose an algorithm to remove foreground occlusions before reconstructing backgrounds. Occlusions are identified by estimating the color variance on elemental images and then deleting it in the final synthetic image. We demonstrate the superiority of our method by presenting experimental results as well as comparing our method with other approaches.

Traditional color-based mean shift tracking algorithm is unable to accurately track the object. To address this problem, we present an improved tracking algorithm. The improved tracker integrates the color and motion cues which characterize the appearance and motion information of the object, respectively. These two visual cues can complement each other and make for more precise target localization. Experiments show that the proposed tracking algorithm has better performance than the traditional mean shift tracker.