Random Coupled Neural Network with Sand Cat Swarm Optimization for Automatic Object Detection in Aerial Images

Object detection in aerial images plays a crucial role in parking space management, traffic monitoring, and surveillance systems. Every day, vast amounts of aerial and satellite imagery are being collected, and analyzing this extensive data can greatly benefit numerous practical applications. However, traditional Automatic Object detection in aerial image methods struggles with challenges posed by complex backgrounds and detecting small objects in large-scale scenes. To address the previously mentioned challenges, a Random Coupled Neural Network with Sand Cat Swarm Optimization (RCNNet+SCSO) is proposed for automatic object detection in aerial images, ensuring high accuracy. In this approach, images from the UC Merced dataset are used, and to initially reduce noise in the input data, Shape-Aware Mesh Normal Filtering (SAMNF) is introduced. Next, the pre-processed images are subjected to object detection using Improved YOLO-V4. Following that, classification is performed with the Random Coupled Neural Network (RCNNet), which is optimized by Sand Cat Swarm Optimization (SCSO), resulting in automatic object detection in aerial images with enhanced efficiency and accuracy. The performance of the proposed RCNNet+SCSO is evaluated using the UC Merced dataset, achieving 99.2% accuracy and demonstrating superior results compared to existing methods. The outcomes of the proposed technique showed automatic object detection in aerial images that enhance the efficiency of various applications, including urban planning, traffic management, and surveillance. By leveraging advanced algorithms, it accurately identifies vehicles, buildings, and other objects, even in complex environments. This technology offers improved scalability and precision compared to traditional methods.