High-Resolution and High-Frame-Rate Image Estimation with Heterogeneous Cameras

In this study, we propose a method of estimating high-frame-rate, high-resolution (HR) images with two cameras having different resolutions and frame rates in real time. Estimation of high-frame-rate HR images is important in many applications of high-speed vision for image recognition at a speed of 1 kHz or greater. In general, high-speed vision is useful for real-time control. However, in conventional systems, the image resolution is often insufficient. Our approach involves estimating unknown HR images from low-resolution (LR), high-frame-rate images which are filtered based on an a priori relation between the HR and LR images. In an experiment, we verified the proposed method based on objective and subjective evaluations.