Infrared micro-scanning error compensation algorithm based on edge location

Abstract

For area-array thermal imaging devices, an essential factor affecting the system imaging quality is the sub-sampling caused by oversized discrete sampling pitch. In order to obtain higher spatial resolution, staring infrared focal plane array (IRFPA) gets multi-frame sub-sampling images by micro-scanning movement to achieve an adequate spatial sampling frequency. However, influenced by external environment and the accuracy of the scanning system itself, the relative displacement between the detector and the scene cannot be absolutely precisely controlled, but exist some error, which will affect the final performance of the reconstructed high-resolution image. We analyzed the distribution of the error and then proposed an infrared micro-scanning error compensation algorithm based on edge location, which is inspired by human retina fixational eye movement pattern. It first locates the edge point in the reconstruction unit and finds the corresponding characteristic values. Later on, matches the characteristic value with the fixed templates and reorders the pixel responses in reconstruction unit utilizing the gray correlation. Finally, it compensates the error real-timely through repeated update and iteration. We apply the algorithm in video sequences acquired by 4-step infrared micro-scanning system. The experiment results show that when aligning to a static scene or stationary region in dynamic scene, the algorithm possesses good resolution enhancement effect, particularly, can improve the clarity and the accuracy of static image edge details.