Gap Image Classification Based on PCA-SVM with Multiple Feature Fusion

The detection of gap size at the threaded connection of the inner diameter of an aviation rocket body is crucial for the performance of the body. Due to the different processing degrees of the inner diameter during processing, there are two working conditions on the inner diameter surface. The image features of the gap under these two working conditions are different. Therefore, before detecting the gap size, it is necessary to distinguish between these two working conditions. This article proposes a multi feature fusion classification method based on PCA-SVM, which takes texture features, grayscale features, and depth features as joint feature parameters of the image. The PCA algorithm is used to perform multi feature fusion on them, and the fused feature parameters are input into the training SVM classifier for training the classification network. The experiment shows that the proposed method has a classification recognition rate of 99.5% in the classification of rough and fine surfaces of aviation rockets, Providing technical support for efficient and automated production of products.