基于LBP-PCA的多传感器目标识别算法

为了更加有效地提高多传感器图像融合后的识别率,提出一种基于LBP-PCA的多传感器目标识别算法。首先分别对红外和可见光图像进行预处理用以突显出要识别的目标,采用LBP算法提取目标的特征点向量,利用PCA算法进行特征融合,得到降维后的融合特征,最后利用SVM(支持向量机)进行分类和识别。实验仿真结果表明多传感器目标经过LBP-PCA融合后在保持足够数量的有效信息基础上降低了特征的维数,有效地提高了目标识别率。

Research on Algorithm of Multi Sensor Target Recognition Based on LBP-PCA

In order to improve the recognition rate of multi sensor image fusion,a multi sensor target recognition algorithm based on LBP-PCA is proposed. First,infrared and visible light images were preprocessed respectively. It is used to highlight the object to be recognized and using PCA algorithm to feature fusion,then the feature of dimension reduction is obtained. Finally,the fusion feature is classified and identified by using SVM (Support Vector Machine). Experimental simulation results show that the multi sensor images based on LBP-PCA fusion can reduce the dimension of the feature and improve the efficiency of target recognition.