Real-time correction of wide-angle lens distortion for images with GPU computing

Abstract

Wide-angle lens provides a broad field of view which benefits some applications, such as video surveillance or endoscopic imaging. However, it also induces lens distortion, especially radial distortion, which may impede further video analysis or perceptual interpretation. For some applications, such as minimally invasive surgery and visual surveillance, realtime correction of image distortion is required. Traditional CPU-centric machines are difficult to achieve the requirement of real-time computation because of a large amount of computation. In this paper, we propose to achieve real-time correction of wide-angle lens distortion of images on several target platforms. In the GPGPU platform, we achieve real-time correction at full-HD resolution by using CUDA. For middle-end devices, an error-controllable mesh is used and the system is implemented by industry standard, OpenGL. We also implement it with OpenGL ES on embedded GPUs for mobile devices. Experiments show using our error-controllable mesh greatly outperforms those using regularly downsampled meshes.

Real-Time Lens Distortion Algorithm on an Edge Device With GPU

“…As a result, they achieved up to 10 frames/second (FPS) at a 640 × 480 resolution with high-end graphics hardware at that time: Silicon Graphics Onyx Reality Engine II. This texture mapping approach has been widely employed for correcting lens distortion in various fields using wideangle lenses, such as HMDs, the medical domain, and surveillance [13], [14]. Traditionally, texture mapping is realized by shader languages like GLSL (OpenGL shader language), and there are two methods: pixel-based and mesh-based [15].…”

“…Lee et al [13] implemented a distortion correction algorithm in a General Purpose GPU (GPGPU) platform with CUDA. They exploited the GPU's hardware-accelerated interpolation ability and achieved a near-real-time performance for HD resolution images (e.g., 48 FPS for 1920 × 1080 on Nvidia GT555M).…”

A Real-Time Hardware Architecture for Image Rectification Using Floating Point Processing

“…, BP(Belief propagation) [2] , AD (Absolute intensity differences) census [3] 교정을 하는 방법 [6,7] 이 연구되고 있으나 현재까지 높은 [26]…”