双目成像算法的研究与FPGA的硬件实现

Binocular stereo vision on computer is based on the mechanism of human eyes. It obtains information from the stereo image pairs and gets the three-dimensional information of the object. With the development of machine vision, binocular stereo vision technology now play a more and more important role in many areas such as terrain mapping, autonomous navigation system and human-computer interaction. These applications catalyze more requirements on the size, the power consumption and the speed of the new generation of stereo vision system. And the implementation of a smaller and more power-saving stereo vision system was no doubt a trend in this area. In recent years, due to the technological advance in the development of Field Programmable Gate Arrays (FPGAs), a single FPGA can now integrated components more than just programmable logics. Embedded processors, digital signal processing (DSP) module, and other specified design unit, can also be seen on FPGA chips now. Meanwhile, as the Electronic Design Automation (EDA) Tools for FPGA design gradually mature, it is also possible that FPGA could be utilized to realize image processing and computer vision algorithms. With higher parallelism and pipeline in computation, solutions on FPGA are no doubt a more optimal choice than those on traditional digital signal processors (DSPs). Moreover, with the widespread of System on a Programmable Chip (SoPC) design, the hardware and software co-design on FPGA will undoubtedly be a good solution for a stereo vision system with smaller size, higher efficiency and lower power consumption.