The Application Of Image Processing Technology In Print Detection System

With the development of printer detection technology, non-contact surface inspection system based on machine vision has become the main form of printed matter detection. It collects a model image as inspection standards, when inspecting it collects images from the production line, and match、correspond、 compare the real-time image to the model image in order to determine whether the current sample defects. Because real sample is transmitted through the belt, so the collected image includes the belt background and the print foreground. Therefore, when generating the model image printed portion of the effective, we need to extract the foreground region and extract the ROI in the effective area of the extracted foreground.The traditional way of ROI extraction is all by-hand. When the model image is generated, the worker should manually draw out the foreground region and the in the drown area, he/she then chooses the right ROI based on his experience. This artificial division method is increasingly becoming a bottleneck of system performance with the detection system hardware upgrades and software system optimization.This essay is based on digital image processing algorithms and uses the algorithms to printer surface inspection system of BJ Focusight. It combines image processing and inspection system, and studies the traditional extraction prospects then make it automatically. This paper contains the following aspects:(1)First, the essay summarizes the development of surface-based inspection system based on machine vision, and studies the existing status of the model image generation method and deficiencies. And then proposes the objectives and requirements of automatic foreground extraction and automatic positioning extraction.(2) Template image generation process in print detection system are analyzed, the demand for image prospects for automatic extraction of generating process is analyzed, combining with the current research progress of print image foreground extraction algorithm and the applicability of the foreground extraction to print images, this paper proposes a printed matter based on Canny edge detection and Hough straight line detection image foreground extraction algorithm, and apply it to the research background of this paper print detection system.(3) Study the characteristics of image, and combine the specific use of nuclear localization methods and requirements. After mainly analyzes the image color feature, texture feature, regional characteristics, point feature, I propose a feature based on the image entropy and SURF points to check the evaluation of possible nuclear positioning area, and use the index of confirmatory to test.(4) Analyze and compare the performance of the algorithm and the comfortable using ways, the algorithm summarizes the advantages and disadvantages, and proposed future research directions.