The work aims to design a defect detection system for packing case type mark based on image processing, in order to realize the real-time and dynamic detection of the type mark defects on the air-conditioning packing case. The hardware platform of the control system was designed based on AM5728 controller, which mainly including the control unit, image acquisition and processing unit, imaging unit, etc., and the actual test research was carried out. Several key methods were adopted, including image enhancement, morphology, defect detection, dynamic threshold segmentation algorithm, etc., and the registration area was selected according to the characteristics of the packing case type mark image. At the same time, a dynamic threshold segmentation algorithm was given to realize defect detection by various algorithms. 250 bar code samples are collected, and the monitoring data are obtained by using the method in this paper. The accuracy rate is as high as 97.2%, and the rate of missing inspection is 0.This method has high reliability and generality and can realize the defect detection of the type mark of the packing case quickly, and solve the practical engineering problem of the real-time dynamic defect detection of the type mark on the air-conditioning packing case.