基于改进混沌差分进化算法的机体损伤区域划分

Airframe Damage Region Division Based on Improved Chaotic Differential Evolution Algorithm

为准确高效地为飞机智能维修提供支持,提出一种基于多维灰度熵理论和改进的混沌差分进化算法的机体损伤区域划分方法。通过分析多维灰度熵的阈值分割原理,将其定义为机体损伤区域划分的适应度函数。引入Logistic混沌模型,并采用循环映射将混沌变量的值域对应至优化变量,改进混沌差分进化算法的寻优过程,提高算法的遍历性。选用不同类型的飞机机体损伤图像进行划分实验。实验结果表明,该方法划分的损伤区域图像清晰有效,且误差率低,与灰度熵穷举法相比,运算速度有明显提升,有效解决了差分进化算法中后期收敛停滞、样本点分布不均匀等问题,能够更好地满足飞机智能维修技术的要求。

In order to support the aircraft intelligent maintenance efficiently, an airframe damage region division method based on multi-dimension gray entropy and improved Chaotic Differential Evolution （CDE） algorithm is proposed. By analyzing the principle of gray entropy threshold division, it is defined as the adaptive function of airframe damage region division method. Then,Logistic chaos model is introduced,and cyclic mapping is used to mapping the range of chaotic variables into optimization variables, the searching optimization process of the chaotic differential evolution algorithm is improved, and the ergodicity of the method is enhanced. Airframe damage region division experiments are performed by different types of damage image. Experimental results show that divided damage region is clear and effective, and error rate is low. Compared with gray entropy exhaustion method, running speed of the proposed method is raised obviously. The problems of differential evolution algorithm such as stagnation of convergence in the middle-later period of the iteration ,uneven distribution of sample points are solved effectively. It can be much better to meet the requirement of aircraft intelligent maintenance.