【摘要】 提出了一种获取正则化参数的新方法。利用随机理论解决正则解模糊误差能量期望值最小化问题,确定正则化参数。对正则化算子给定为Laplacian算子的情形予以测试,实验结果表明该文的恢复技术比传统方法的恢复性能好,恢复效果接近最佳且性能稳定。更多还原

【Abstract】 A new technique for selection of regularization parameter is proposed in this paper.The goal is the mini-mization of the blurred error of regularized image.Using stochastic theory,the new method minimizes the expectation of the energy of the blurred error of regularized image,and then determines the regularization parameter.In the case of the regularization operator with Laplacian,numerical results show that the new method performs better than traditional meth-ods and yields steadily close-to-optimal restoration.更多还原