A Portable Single-Pixel Object of Interest Imaging Camera

Single-pixel cameras are generally large, stationary systems designed for laboratory use. Their imaging process relies on reconstructing entire scenes by processing the entirety of the collected data through specialized algorithms. This method, while functional, is inherently data-intensive and lacks efficiency. This paper presents an innovative portable single-pixel object imaging camera that combines single-pixel camera with artificial intelligence to selectively image specific objects from complex scenes. By leveraging compressed sensing and neural network optimization principles, this method achieves efficient imaging while significantly minimizing data redundancy and transmission artifacts, all integrated into a portable medium. Built upon the foundation of traditional single-pixel cameras, this Portable camera is the first to actively achieve imaging of objects of interest, enabling real-time tracking of targets anywhere. Its potential applications span multiple fields, including autonomous driving, remote sensing, medical imaging, and intelligent robotics.