Developed a robust, real-time human detection system for UAVs for rescue operations in challenging environments, such as natural calamities. Used multi-model solution leveraging YOLOv8 for accurate human detection. In this YOLOV8 is not directly used but there are changes made to the YOLOV8 backbone architecture as mentioned in the refernce paper in order to enhance object detection for UAVs. Advanced Pre-Processing done in order for enhanced image qulaity for better detection. Dehazing and de-blurring techniques also used to enhance detection accuracy in smoky or hazy conditions. Applied a CNN-based model to upscale low-resolution images, improving detection of small human figures captured from high altitudes by UAVs.
Shravyabhat/Human-Detection-in-Hazardous-Situations-with-UAVs-
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|