基于曲线拟合的块匹配法应用于FBG图像解调系统

Application Of Block Matching Based On Curve Fitting To The Embedded Fiber Bragg Grating Image Demodulate System

随着嵌入式技术的迅猛发展,利用嵌入式处理器构建FBG图像解调系统实现大量传感数据的实时、实地快速处理成为可能。受运动估计算法中块匹配法的启发,考虑到质心法的不足,本文将曲线拟合的块匹配法应用于嵌入式FBG图像解调系统,测量FBG光斑的位移值。实验结果表明,目标经过块匹配处理后,用二次曲线拟合绝对差和(SAD)值,得到的最小块误差(MBD)点具有亚像素精度定位,用该算法对FBG解调系统解调,具有稳定性好、精度高、收敛速度快的优点,解调结果十分理想,适合应用于嵌入式FBG传感系统的光斑图像解调。

With the fast development of embedded technology, the Fiber Bragg Grating （FBG） image-demodulated system designed through the embedded microprocessor make it possible to implement real-time and fast processing of a large number of sensing data. Inspired by the Block Matching based on Motion Estimation Algorithm, considering the disadvantage of Centroiding Algorithm, the Block Matching is applied based on Curve Fitting to the embedded Fiber Bragg Grating image-demodulated system in order to measure the displace of FBG spot. The experimental results show that the MBD point can be calculated more precisely by fitting SAD and the displacement of the light spot can be located with sub pixel accuracy after target blocks are processing by block matching,. This method applying to the embedded Fiber Bragg Grating image demodulate system displays improved stability, highaccuracy and fast convergence. The demodulate results are very satisfactory in the embedded Fiber Bragg Grating image demodulate system.