Research And Implementation Of Key Algorithm Of Shadow Correction In Video/Image Processing Based On FPGA

With substantial increase of video image data bandwidth, FPGA-based video image processing method is increasingly demonstrated its high-performance. Because of the parallel mechanism, it is good at processing real-time video image at a lower clock frequency based on FPGA. This paper studies the design and implementation of medical image shadow correction algorithm based on FPGA.Firstly, this paper analyzes the cause of image shadow, discuss the principles of several correction algorithms, and choose the parametric shadow correction for medical X-ray image processing. In order to implement the algorithm in FPGA, the implementation of square root algorithm must be solved.Secondly, this paper introduces three major square root algorithms: digital recurrence algorithm (including restoring algorithm, non-redundant algorithm and SRT algorithm), Newton-Raphson algorithm and look-up table algorithm. The paper focuses on the principle of digital recurrence algorithm, and compares the performance of several algorithms.Thirdly, this paper studies the implementation of SRT algorithm in FPGA, and presents resolutions to several key problems encountered, including the choice of radix, the form of square root, the form of partial remainder, the conversion of square root, the digital selection rules, the initialization and termination conditions, and the results amendment. Then the algorithm is described in Verilog HDL.Finally, the paper synthesizes and simulates the codes of square root algorithm, and analyzes the simulation results. Then, the paper summarizes the major work and suggests the future work.