An Image Stitching Technique Based on Image Moments
Changjin Li,Jiquan Yuan,Xinyan Cao,Wei Ren,Xixin Cao
DOI: https://doi.org/10.1109/icicr61203.2024.00028
2024-01-01
Abstract:With the evolution of technology and the advancement of computer science, image moments have gained increasingly widespread application in the field of image analysis. Geometric moment computation stands as a pivotal topic in the domain of image processing. Nevertheless, the computational complexity associated with moment calculations is relatively high, to a certain extent constraining its practical applicability. This paper delves into the investigation of various aspects, including the representation of high-order moments, the selection of image moment orders, parallel computation of image moments, and image stitching based on image moments. It deduces high-order moment functions, utilizing these higher-order moments for the extraction of image features, thereby enhancing the accuracy and stability of image feature representation. Based on actual images, the algorithm selects the expression of moment function orders, balancing the choice of image moment orders by weighing preset target quantities and sampling parameters. This provides a more targeted image moment calculation scheme for practical applications. The moment calculation algorithm is optimized by transforming multiplication operations into addition operations, simplifying the complexity of the original serial calculations. Through parallel computation, the computational workload is reduced, thus enhancing the performance of image moments in practical applications. Employing machine learning methods and deep learning techniques in conjunction with image moments facilitates image recognition and classification, thereby improving recognition efficiency.