A Method for Traffic Signs Recognition Using PCA and LDA

摘要

In this paper, a novel method for traffic signs recognition based on PCA and LDA is proposed. To enlarge the classification distance between two different traffic signs samples, normalization of within-class means is considered firstly. Then the eigen space of all samples is calculated using the Principal Component Analysis (PCA) method and Linear Discriminant Analysis (LDA) method, respectively. Next, the above two eigen spaces is mixed to produce the best classification space. Then the training and testing samples are projected into the mixtur feature space to get their features, respectively. The Nearest Neighbor Distance (NND) method is employed as a classifier for traffic signs discrimination. The performance of the proposed method is validated with a validation test database consisting of 40 regulatory traffic signs image. Experimental results demonstrate that the proposed method is efficient and effective.