A new stereo matching hardware with merged odd-even PE architecture

In this paper, we propose the new hardware architecture which implements the stereo matching algorithm using the dynamic programming method. The dynamic programming method is used in finding the corresponding pixels between the left image and the right image. The proposed MOEPE (merged odd-even PE) architecture operates in the systolic manner and finds the disparities from the intensities of the pixels on the epipolar line. The number of PEs used in the MOEPE architecture is the number of the range constraints, which reduced the number of the necessary PEs dramatically compared to the traditional method that uses the PEs with the number of pixels on the epipolar line. For the normal sized images, the number of PEs in the MOEPE architecture is less than that of PEs in the traditional method by 25 times. The proposed stereo matching hardware called MOEPE consists of three parts, minimum cost module which finds the minimum matching cost and path generator which finds the minimum cost path and disparity extractor which extracts the disparities. The proposed architecture is modeled with the VHDL code and synthesized by the SYNOPSYS tool. The stereo matching hardware, MOEPE, with 9 PEs is designed with 33,660 gates in SAMSUNG SOG process.