Hemispherical 3D Around View Monitoring Algorithm Using Image Synthesis of Multi-Channel Cameras

Abstract

Most current vehicle black boxes with AVM (Around View Monitoring) function provide a bird's-eye view of 2D type. This viewpoint is a form of looking at the vehicle from above so that it gives the driver a sense of spatial heterogeneity. In addition, the top-view type has a limited viewing angle, which makes it difficult to conduct an accident investigation, parking and so on. To solve these problems, we propose the 3D A VM algorithm using image composition of multi-channel cameras in this paper. The proposed 3D AVM algorithm is divided into three stages. In the first stage, we perform distortion correction of wide-angle lens and perspective transform. The image stitching, white balance and alpha blending are made in the second stage. The 3D texture mapping is executed in the final stage. This proposed mapping method, which combines planar and hemispherical projection method, makes it possible to see around the vehicle at various viewpoints of 360-degree. The experimental study results indicate that the proposed method overcome the disadvantages of the 2D AVM black boxes and may be useful for safe driving and accurate accident investigation.