Search

Aiming at the problems like many locations to be detected and complex texture background for automobile calipers, a detection method combining gray average algorithm, wavelet denoising and LBP operator is proposed. First, the region of interest is extracted from the caliper sample and the wavelet denoising is preprocessed. The gray-to-average ratio of the genuine template and the sample at each position is calculated, then a suitable threshold is set to remove the defective maps with large differences of the ratios. ff the gray-to-average ratio is within the indistinguishable range M, the feature is extracted by the LBP operator. Finally, the template feature is compared with the image feature to be detected in the M range to measure the caliper defect. The experimental results show that the proposed method has real-time characteristics and reliability for the detection of automobile calipers, which can be well applied to the detection of actual products. [ABSTRACT FROM AUTHOR]