Image Registration of High Resolution Chandrayaan-2 Data

Accurate and reliable detection of features, along with effective mapping and registration, has been a topic of considerable discussion. Detecting and identifying craters on the lunar surface poses significant challenges due to subtle intensity variations and shadow effects. Conventional corner and edge detection methods often struggle to outline these craters, complicating the image registration. This paper proposes a novel strategy for data preprocessing to enhance edge and corner detection for OHRC and NAC datasets. Utilizing SIFT (Scale-Invariant Feature Transform) and ASIFT (Affine Scale-Invariant Feature Transform) algorithms, the study identifies and maps key points between the OHRC (data to be registered) and NAC (reference) datasets, achieving high pixel-level registration accuracy (within the range of 0.2-2.5 pixels). The application of image stitching and warping, results in georeferenced images that are both accurate and precise, contributing to improved navigation and enhanced image quality.