Research on engineering application of deep learning model based on UNet structure in infrared weak target detection

Infrared weak target detection has an important application value in the fields of early warning, traffic management and security. However, limited by the weak characteristics and complex background of the target, the traditional detection methods are insufficient in terms of accuracy and real-time performance. In this paper, we propose a deep learning algorithm based on the Transformer structure with an improved Unet model for efficient detection of infrared weak targets. In order to realize engineering deployment, we deploy the model on two embedded platforms, RK3588 and Nvidia Orin Nano, and conduct performance comparison experiments. The experimental results show that the proposed method outperforms the traditional method in terms of detection accuracy and real-time performance, and exhibits good adaptability and high efficiency on different hardware platforms.