Abstract:A system of determining acrylamide contents in fried potato chips based on computer vision technology was developed. For each 60s, the image acquisition device collected double-sided images of potato chips removed from the fryer. The acrylamide contents of the corresponding potato chips were determined by gas chromatography-mass spectrometry (GC—MS) method, meanwhile, the target potato chips of the double-sided images were extracted and L*,a*,b* average of the overall target was measured by proposed image filtering algorithm, the color image segmentation algorithm and the color measurement algorithm, respectively. The obtained data was analyzed and a strong correlation between the average of double-sided a* average and acrylamide contents of potato chips was found (R2=0.971) and a linear regression equation was established. Potato chips double-sided images of 10 different brands sold on the market was collected and the average of double-sided a* average was put into the equation. The maximum relative error between the acrylamide contents calculated and standard chemical method was 4.94%, which indicated that the method was feasible and accurate.