In view of the problem of low accuracy and low efficiency of manual verification for the recycling classification of old smart meters removal, this paper proposes a method of parameter information detection based on machine vision. By detecting the rated parameter information of smart meters, the classified recycling work of electricity meter is completed. Firstly, based on the establishment of a smart meter image detection system using c # and Halcon as software platforms, the Blob analysis algorithm is used to first extract the ROI (region of interest) of the image, and the histogram equalization is used to process the extracted image to enhance the contrast between the background and the target area, and obtain a high-quality meter image. The Canny edge detection algorithm is improved through the OTSU algorithm, so as to improve the adaptability of the image threshold range and obtain a more complete image appearance contour. Finally, the character segmentation processing is performed for image to obtain the rated parameter information of electricity meters. After experimental verification, this method can accurately detect the rated parameter information of the identification plate of the electricity meter. The experimental data shows that the detection accuracy rate is 99.5%, and the average test time of each meter is 0.62 s, which greatly saves the verification time of the meter classification work, and improves the efficiency and accuracy of the work.