Weld bead geometry inspection by non-destructive testing techniques is the major challenge of today's welding industries. As compared to similar materials, application of dissimilar materials demanded all over. Generally, feature extraction by radiographic images sensed geometric features and categories for classification of weld imperfections. Whereas diverse understanding for human intelligence grasp beneficial evidence from non-geometric features of images. To overcome this difficulty by exploring features and recognize imperfections of two-dimensional digital image wherever geometric features state presence of weld. Mostly, classification accuracy significantly influences by weld imperfection region segmentation and imperfection texture feature extraction. The proposed techniques of local binary pattern in which local binary code describing region, generating by multiplying threshold with specified weight to conforming pixel and summing up by grey-level co-occurrence matrix to extract statistical texture features. At last, support vector machine and k-nearest neighbours compared to discover finest classifier and uppermost accuracy of 96% attained through grouping of local binary pattern features and support vector machine.