Abstract: In order to realize the automatic and intelligent reading of the precipitation value of Sodium Dodecyl Sulfate（SDS）, an automatic image recognition system was designed in a shaking machine for SDS precipitate value automatic determination. After the feasibility analysis of three image processing methods, an image recognition method based on label comparison was selected. This method included two modules such as image preprocessing and volume value conversion. The main steps included converting RGB to HSV, multi-channel binarization, extracting label and precipitate value foreground, finding label and precipitate value contour, etc. It could process multiple test tubes at the same time and give the volume of each tube. The experimental results showed that the average recognition accuracy was 98.945%. Since the image processing light source was fixed, compared with manual reading of indicator values, it not only improved the experimental efficiency, but also reduced the subjective human interference caused by different lines of sight each time reading indicator values, effectively improved the accuracy and realized the automation and intelligence of reading precipitation values.