Orthomosaicking Thermal Drone Images of Forests via Simultaneously Acquired RGB Images

Operational forest monitoring often requires fine-detail information in the form of an orthomosaic, created by stitching overlapping nadir images captured aerial platforms such as drones. RGB drone sensors are commonly used for low-cost, high-resolution imaging that is conducive to effective orthomosaicking, but only capture visible light. Thermal sensors, on other hand, long-wave infrared radiation, which useful early pest detection among applications. However, these lower-resolution suffer from reduced contrast and lack descriptive features successful leading gaps or swirling artifacts orthomosaic. To tackle this, we propose a thermal orthomosaicking workflow leverages simultaneously acquired images. The latter producing surface mesh via structure motion, while texture this yield Prior texturing, RGB-thermal image pairs co-registered using affine transformation derived machine learning technique. On average, individual achieve mutual 0.2787 after co-registration our technique, compared 0.0591 before co-registration, 0.1934 manual co-registration. We show orthomosaic generated (1) better quality than existing methods, (2) geometrically aligned with (3) preserves radiometric (i.e., temperatures) original imagery, (4) enables easy transfer downstream tasks—such tree crown also provide open-source tool implements facilitate usage further development.