Aerial Image Detection Algorithm Based on Improved YOLOV8s

With the rapid development of drone technology, target detection based on drones has become a hot spot in current research. However, existing methods face challenges such as low detection accuracy and poor real-time performance. To address this problem, an improved target detection method YOLOV8s-SME is proposed for application in aerial images. This algorithm combines the lightweight StarNet backbone network, replaces the traditional feature pyramid network, and uses the multi-scale feature fusion module (MFFM). In addition, EUCB's upsampling strategy and Dsmehead are used to improve detection accuracy. Experimental results show that on the Visdrone2019 data set, YOLOV8s-SEM significantly reduces the amount of parameters and calculations while maintaining detection accuracy. The inference speed and post-processing time are superior to traditional methods, and is suitable for drones with limited computing resources. Real-time target detection applications on the device.