A new fast and strong robust recognition algorithm for strain gauge

Strain gauge´s fast and effective recognition places a great importance in industrial applications. In actual application, the absorption position of the strain gauge is not unity and the strain gauge is quite light and thin, which cause the absorption of the strain gauge of the mechanical arm produce much revolves and translates. The traditional Wavelet moment algorithm can well reflect the regional information of image and is invariant to rotation, scale and offset, but for the industrial production, the operand is big, which is not suitable for the fast request. To address this shortcoming, this paper uses the multi-criteria decompose of the wavelet under the polar coordinate algorithm to extract wavelet moment of the low frequency sub graph, then uses PCA to do the recognition. Experimental results obtained by our methods demonstrated that the wavelet moment combined with PCA algorithm can obtain 100 percent of the recognition rate and c an massively reduce the strain gauge´s recognition time, which can be well used in industrial applications.