Abstract: In order to solve the problem of carton missing detection after cigarette sticks are packed, a carton missing detection system based on millimeter-wave radar imaging and computer vision was proposed. Firstly,millimeter-wave radar was used to collect the echo data of cigarette sticks in the cigarette box. Secondly, the synthetic aperture radar imaging algorithm was used to realize the two-dimensional image reconstruction of the cigarette box echo data. Finally, the intelligent identification of the carton- missing was realized by collecting a large number of pictures of carton-missing and training the defect samples by using computer vision technology. The defect rate of inline detection system was higher than 99.9%, and falsealarm rate was lower than 0.033% through long-term observation and sampling,and the detection rate and false alarm rate met the requirements of real-time detection. In addition, the method is suitable for different tobacco specifications.