| Popis: |
In order to realize burning inspection for the train wheel tread, a burning dynamic inspecting technology using vision were investigated. Firstly, the tread region was segmented by tread segmentation technology based on improved Canny algorithm combined stationary wavelet edge detection algorithm. Then the extracting technologies of the suspicious regions of the wheel tread burnings based on tread boundary searching technology are studied. Finally, GA-BP algorithm was applied in the pattern recognition of burnings based on the feature extraction. The experimental results show that that the burning effective detecion was 96.67%, the fulfillment rate was 65.7%.Therefore, the proposed method was effective method of burning dynamic inspection. |
