某型飞机红外辐射特征聚类算法

Research on Clustering Algorithm of Aircraft Infrared Radiation Characteristics

研究某型飞机的飞行状态对其红外辐射特征的影响,针对飞机的红外辐射特征对飞行状态的改变比较敏感,需进行准确地分类来判断飞机的飞行状态。传统的K均值算法易陷入局部最优解,导致了判断飞行状态的正确率低。因此,提出一种改进的K均值算法通过对飞机的红外辐射特征有影响的因子来进行聚类,分析判断出飞机的飞行状态。该算法将遗传算法的全局搜索能力快与模拟退火算法的局部搜索能力强的优点相结合,从而避免了K均值算法陷入局部最优解,仿真表明该算法较原算法有更好的寻优能力,能够准确地对飞机的红外辐射特征聚类,是一种有效的算法。

To study the influence of a certain type of aircraft flight status on the infrared radiation characteristics,aiming to sensitivity of the changing on the infrared radiation characteristics to the aircraft flight state ,it is necessary to classify the characteristic data clearly to estimate the flight status. The traditional K-means clustering algorithm has the disadvantage of weakness in overall search, easily failing into local optimization ,which will lead to the low correct rate of judgment of flight status. So an improved K-means algorithm is put forward to analyse and decide them by clustering the influential factor of infrared radiation characteristics. The improved algorithm combines the advantages of global search ability of GA and local search, avoids K-means algorithm to lost into local optimal solution. The results of simulation show that the performance of above-mentioned algorithm is better in the optimization capacity than before,and is accurate in clustering the infrared radiation characteristics. At last, it is an effective algorithm.