A novel object localization encryption algorithm for wind turbine

Currently, wind turbines are an efficient and environmentally friendly wind power generation technology. Given the aging and wear of wind turbines, it is necessary to conduct regular inspections using unmanned aerial vehicles. However, aerial images may leak critical information. To protect the security of aerial images, this paper utilizes the YOLOv5 model to locate wind turbines in aerial images and proposes a new chaotic system based on the combination of sine and cosine functions, called the tent-dynamic cross-coupled mapping lattice chaotic system. Experimental analysis demonstrates that this chaotic system exhibits better chaotic characteristics compared to the original tent-dynamic cross-coupled mapping lattice. Furthermore, a fast scrambling image encryption algorithm with a dual diffusion model is proposed, utilizing a chaotic matrix during the confusion phase to rapidly perturb the original image. During the diffusion phase, a single diffusion model is achieved by randomly transforming one pixel of the original image to affect all pixels, followed by a dual diagonal diffusion of image pixels. Experiments show that this algorithm has advantages such as fast execution speed, large key space, robustness, and high security.