Estimating Actual Size of Missile on Air using Object Detection Algorithm

Missile size measurement devices find application across diverse sectors including research-development for designing and verifying prototypes, military bases, and in space exploration. In this project, we achieved significant progress in implementing an instance segmentation pipeline for missiles using YOLO. We collected datasets, preprocessing with annotated images, segmentation masks, training a YOLO model for instance segmentation, performing inference on images to obtain precise segmentation masks for detected missiles. By calculating the pixel area within these masks, we estimated the size of each missile. We evaluated the model's performance using segmentation-specific metrics, ensuring accuracy. Future steps involve fine-tuning, and ongoing monitoring.