Research Of Image Recognition And Detection For Hot Stamping Area Based On CCD

Along with the progress of the society and the improvement of people’s living standard, people put forward higher requirements for packaging decoration printing. Because of its unique metallic luster and strong visual result, as an important printing surface finishing processing technology, Hot Stamping Process has a wide range of applications. However according to quality problems of hot stamping process, at present quite part of enterprises are still using visual and subjective judgment of the person. Considering the situation of research and application of this technology, this paper focused on the recognition and detection technology of hot stamping imageTo meet printing requirement of real-time online detection, this paper proposes a adaptive image recognition algorithm based on projection characteristics. Firstly, median filtering method was applied to eliminate noise. Then, according to the chrominance threshold of hot stamping image, we can precede threshold segmentation. Finally, according to the coordinates of horizontal projection and vertical projection, we can determine the precise location of the hot stamping image in the whole picture. The experimental results show that the algorithm can locate the exact location of hot stamping image, high matching speed, and strong anti-jamming capability.Considering the speed and accuracy of on-line detection, we adopt colorimetric assays to judge the defects of hot stamping image, which has realized the defect image segmentation on the image, and combining with the algorithm of Blob, judgment of defect area, find the connected domain, and to determine the number of connected domain, such as size, area, location information.During the study, Matlab image processing programming was used to research image recognition algorithm. Finally, the hot stamping image identification test system of print whole image was built by Microsoft Visual C++. The experiment results showed the image recognition and detection methods proposed can realize image defect marking, and classification accurately in hot stamping image quality detection process, and they had good real-time performance and good serviceability.