摘要
文中提出一种基于量子行为粒子群优化(QPSO)的图像边缘检测算法。该算法将4个自适应神经模糊推理系统(ANFIS)子检测器和一个后处理块组成一个图像边缘检测器。每个ANFIS子检测器都是一个4输入单输出的一阶Sugeno模糊推理系统。ANFIS中的前提参数用QPSO优化,结论参数用线性最小二乘法优化。新算法特色在于能有效地提取噪声图像中的边缘信息而无需进行图像滤波预处理过程。仿真实验结果表明,新算法提取边缘信息的能力明显优于传统的边缘检测算法。
An image edge detector based on the quantum-behaved particle swarm optimization （QPSO） is presented. The edge detector is constructed by combining four ANFIS with a postprocessor. Each ANFIS is a first-order Sugeno type fuzzy inference system with 4-inputs and 1-output. The antecedent parameters of each ANFIS subdetector are optimized by using QPSO and the least square estimation method is employed to determine the consequent parameters. The advantage of the proposed algorithm is that its efficient extraction of edges in gray images corrupted by impulse noise without a noise filter prior to edge detection. The experimental results show that the new algorithm significantly outperforms other conventional edge detectors.
出处
《江南大学学报（自然科学版）》
CAS
2015年第2期127-135,共9页
Joural of Jiangnan University (Natural Science Edition)
基金
国家自然科学基金项目(61170119)
中央高校基本科研业务费专项项目(JUSRP211A38)
