Research On Printed Image Detection Technology

In the field of printing image detection technology,visual technology with the traditional artificial has incomparable advantages.The subject mainly talks about the image edge detection,operators and the improved image feature matching,completed the design of the process for defect inspection of printing image,and mark the printing defect image morphology.Subject of experiment for visual inspection technology provides a certain reference for eliminating the subjectivity of artificial detection,and has a great significance for improving the detection speed and precision.Subject combined with the principle of image processing and printing,through the comparison with experimental data analysis,studied the method of printing image detection algorithm for the theoretical analysis,optimizing improvement and experimental research.Preprocessing the printed image,detecting the edge of printed image,and improving edge detection operator so as to obtain a good result.Through contrast experimental result and experimental data analysis,the method of establishing detection speed is improved and the improved sobel operator edge detection is more complete;Then in the scale invariant feature transform algorithm,the acquired printed image is extracted with the standard image by the Guassian difference pyramid to complete the feature matching,and realized the distinction of printing defect image;Finally applied the improved sobel operator on the edge of the standard printed image and detection image,after all processing printing defects could be flagged,and completed printing image morphological defects detection.The results show that the adaptive median filtering processing printing image effect is better;Improving detecting speed,sobel operator edge detection is more complete;Through the scale invariant feature transform algorithm can realize the match of the defective images;The inspection process can mark defects correctly.