Abstract:Domain of stereo vision is highly important in the fields of autonomous cars, video tolling, robotics, and aerial surveys. The specific feature of this domain is that we should handle not only the pixel-by-pixel 2D processing in one image but also the 3D processing for depth estimation by comparing information about a scene from several images with different perspectives. This feature brings challenges to memory resource utilization, because an extra dimension of data has to be buffered. Due to the memory limitation, few of previous stereo vision implementations provide both accurate and high-speed processing for high-resolution images at the same time. To achieve domain-specific acceleration for stereo vision, the memory limitation has to be addressed. This article uses a Mini-Census ADaptive Support Region (MCADSR) stereo matching algorithm as a case study due to its high accuracy and representative operations in this domain. To relieve the memory limitation and achieve high-speed processing, the article proposes several efficient optimization methods including vertical-first cost aggregation, hybrid parallel processing, and hardware-friendly integral image. The article also presents a customizable system which provides both accurate and high-speed stereo matching for high-resolution images. The benefits of applying the optimization methods to the system are highlighted. With the aforesaid optimization and specific customization implemented on FPGA, the demonstrated system can process 47.6 fps (frames per second) and 129 fps for video size of 1920 × 1080 with a large disparity range of 256 and 1024 × 768 with a disparity range of 128, respectively. Our results are up to 1.64 times better than previous work in terms of Million Disparity Estimation per second (MDE/s). For accuracy, the 7.65&percnt; overall average error rate outperforms current work which can provide real-time processing with this high-resolution and large disparity range.

Analysis of adaptive support-weight based stereo matching for hardware realization

Real-time Stereo Vision System Using Adaptive Weight Cost Aggregation Approach

PC Based Real-Time Implementation for Regional Matching in Stereo Vision

Virtual support window for adaptive-weight stereo matching.

Accurate Real-Time Stereo Correspondence Using Intra- and Inter-Scanline Optimization

Fast Stereo Matching Algorithm Based on Adaptive Window

Dynamic Limited-Search Matching Algorithm Based on Geometric Properties of Stereo Imaging

Hardware Friendly Algorithm of Hr Real Time Stereo Matehing for Automatie Drive

Hardware Friendly Algorithm of HR Real Time Stereo Matching for Automatic Drive

Research and implementation of adaptive stereo matching algorithm based on ZYNQ

Hybrid stereo matching by dynamic programming with enhanced cost entry for real-time depth generation

A Hardware Friendly Stereo Match Refinement Algorithm Using Disparity Gradient Based Region Growth Method

Adaptive Disparity Computation Using Local and Non-Local Cost Aggregations

An Efficient Stereo Disparity Estimation Method Based on Region Information

Stereo matching algorithm using improved guided filtering

A Configurable Local-Expansion-Move Accelerator for Accurate Stereo Matching

Hardware Acceleration for an Accurate Stereo Vision System Using Mini-Census Adaptive Support Region

A Temporally Streamlined Optimization Method for Stereo Video Correspondence

An Efficient Adaptive Window Based Disparity Map Computation Algorithm by Dense Two Frame Stereo Correspondence

Sparse LIDAR Measurement Fusion with Joint Updating Cost for Fast Stereo Matching

Improved real-time three-dimensional stereo matching with local consistency