Study on barcode digital recognition of smart meters based on projection-histogram matching

In order to solve the key technical problem that is barcode recognition of smart meters in smart meters automation detection based on machine vision, the paper presents barcode digital recognition approach based on projection-histogram matching. First of all, locate barcode digital region with Sobel operator, using wavelet transform to enhance the digital region image. And then proceed single character segmentation in the digital character region, in addition normalize the size of character via quadratic interpolation method. The next do projection of horizontal and vertical direction of single character and make statistics of the histogram distribution. The final step is to match the projection histogram and template histogram to identify the corresponding number to get the barcode information of smart meters. The experimental results show that this method can accurately identify the barcode numbers on smart meters, has high recognition efficiency and good stability. Thus the method has high practical value.