An Effective Image Mosaicking Approach for Mid-Wave Infrared Images Based on 2D Log-Gabor Feature Transform and Priori Coordinates

Mid-wave infrared (MWIR) images have been widely used to embody abundant thermal radiation geographic information. However, due to the small field of view (FOV) of MWIR imaging detectors, image mosaicking is necessary to combine multiple images with overlapping regions into a larger FOV image. Feature detection and mosaicking of MWIR images remain challenging due to their low contrast and low signal-to-noise ratio (SNR). To address this problem, we propose an effective image mosaicking method for MWIR images based on 2D Log-Gabor feature transform (LGFT) and priori coordinates. The 2D LGFT based on phase coherence used in the proposed method enables feature point detection independent of image intensity and gradient, from which more distinctive features can be extracted. Priori coordinates are used for feature point matching constraints, and topological analysis is applied for globally consistent alignment to obtain the final mosaic of MWIR images. We evaluate the proposed method on two captured MWIR image datasets under different scenarios, and the results show that our method outperforms other state-of-the-art feature extraction and matching methods, while preserving details and reducing distortion compared to commercial tools.