将指数修正嵌入到矢量外推框架中, 构建了一种改进的矢量外推Richardson-Lucy算法。以结构相似度因子作为评价指标, 优化彩色图像三通道最佳迭代次数,确定矢量外推加速参量与指数修正加速参量最优组合.试验结果表明: 改进的矢量外推Richardson-Lucy算法总迭代次数减少了78.9%, 能有效抑制噪声放大, 无需后续处理就能获得视觉质量更好的复原效果.该研究可为Richardson-Lucy算法在实时波前编码成像中的应用提供参考.

An improved vector extrapolation based on Richardson-Lucy algorithm was designed by embedding the modified exponent into the vector extrapolation. The structural similarity index was used as a criterion to determine the optimum iterations and optimum combinations of two acceleration factors. Experimental results show that total iterations are reduced approximately 78.9% and visually satisfactory restoration results can be obtained without denoising the restored image further. This work provides a reference for the development of the Richardson-Lucy algorithm in the application of real-time wave-front coded imaging.