ERCM: Bionic Event-based Registration Method Based on Contrast Minimum for Intelligent Unmanned Systems

Abstract

Event cameras have the advantages of high time resolution and high dynamic range, and the fusion of its image and RGB image can bring great benefits to intelligent unmanned systems. However, the two modal images are not fully registered, which may affect the performance of tasks in intelligent systems, such as semantic segmentation. In this paper, we propose a registration method of event and RGB cameras based on contrast minimum. Our experimental data consists of 8 sequences, each consisting of 20 consecutive frames from the DSEC dataset, an event-RGB image dataset. As a result, the overlaid image of event images and RGB images show better consistency in visualization. Compared with the baseline, the declining value of contrast is from 26 to 161, the proportion of overlapping pixels in edge and event image has increased by 2.00% and 0.91%, respectively. The study in this paper demonstrates promising applications of the fusion of event and RGB cameras in intelligent unmanned systems.