On image registration and defect detection techniques in the print quality detection of cigarette wrapper

In a print quality detection system, image registration and defect detection algorithms are two key techniques. The precision and speed of them have a direct impact on the performance of the system. According to the demands and characteristics of the print quality detection of cigarette wrapper, an image registration algorithm and an defect detection algorithm in the system were detailedly studied in this paper. The paper proposed an image registration method using a cross-correlation algorithm and a control-point based search strategy and an image defeat detection method combining a dynamic threshold image subtraction technique with a hierarchical detection strategy. The experimental results showed that these algorithms could detect rapidly and effectively the print quality of cigarette wrapper under the condition of the acquired image containing some noises, local defeats and gray-level variations.