Intelligent Image Stitching Using AIML for Eco-Surveillance and Environmental Change Detection

– Environmental surveillance and monitoring involve the continuous observation of a wide geographical region to determine vegetation health, land use changes, and ecological transitions. However, the imaging of wide fields using drones or ground cameras often leads to the generation of overlapping images based on hardware constraints and limited field of view. Manual processing of these images for stitching is a time-consuming process and often involves geometric distortions. This paper proposes a computer vision and Artificial Intelligence hybrid approach for the automated stitching of panoramic images and environmental transition analysis. The image stitching process involves the use of Oriented FAST and Rotated BRIEF (ORB) for keypoint detection and description, followed by feature matching using Brute-Force matching and Random Sample Consensus (RANSAC) for outlier removal. Homography calculation is carried out for geometric alignment of overlapping images, and image warping and alpha blending are used to generate panoramic images with minimal distortions. For improved environmental analysis, the stitched panoramic images are further analyzed using AI-based analysis tools for vegetation pattern analysis and temporal transition analysis.