Study Of Key Technologies On Automatic Optical Inspection System Using Contact Image Sensor

With the development of industrialization and information technology, modern manufacturing demand for products AOI(Automatic Optical Inspection) system increases rapidly. For the detection of large-size and high-precision products, linear array scanning AOI system is very important. The linear array scanning AOI system has become the core and standard industry equipment in the precision electronics, semiconductor, glass, textile, paper, high-tech manufacturing industry and the basic manufacturing. Most of the linear array scanning AOI is mainly industrial camera CCD linear array as the imaging system. In recent years, due to the CIS imaging system with low power consumption, simple structure, compact volume, and low cost, the domestic and foreign present to develop in the direction of linear array scanning AOI based on CIS.With the rapid development of Chinese science and technology industry, industrial restructuring and upgrading is imperative. To achieve the intelligent manufacture of China, the independent of the core intellectual-made equipment is imperative. Linear array scanning and other high-end equipment localization level is far from enough, it is a must to carry out in-depth and systematic research in the key techniques in system design to shorten the gap with foreign leaders.In this dissertation, based on solving key technical issues of linear array scanning AOI system, analysis and experiment are made in the following aspects, which accumulating the corresponding engineering experience and contributing the power for the sustainable development of domestic CIS AOI equipment.1. CIS linear array camera designAccording to the linear array scanning AOI system for imaging system, the key technologies of CIS linear array camera design is researched. Framework based on FPGA platform, through internal and external memory pipelined multiplexing model of flexible, fast data storage and transmission mode, the design and implementation of large data, high rate of CIS linear array camera. According to the test results, the camera at 300 dpi resolution scanning speed is 140 IPM, the delay is less than 1 ms, the peak signal to noise ratio up to 86.9dB before and after the compression of two images. Analysis of CIS influence of exposure time for imaging quality, puts forward the calculation method of exposure time parameters. Exposure times for the same problem, analysis principle of image brightness difference frequency signal contact issue different causes, through the use of two trigger method improved CIS linear array camera internal exposure mechanism to solve the problem. Data show that, the above technology application design of CIS linear array camera scanning speed and scanning results completely satisfy the automatic optical inspection system of high-end and high rate requirements.2. Image correction technology of CISBy analyzing and discussing the cause of non-uniform imaging about CIS linear array camera, the two-point correction method which suits hardware system optimization algorithm is proposed based on the gray two-point correction method. Theory proves that the error rang is less than 0.5 gray compared optimized algorithm with optimized two-point correction. Because the gray often expressed as an integer rang 0 to 255, so the two kinds of methods can achieve the same effect at improving the quality of image. The optimized algorithm because of algorithm is relatively simple, more suitable for the realization on hardware platform. In order to further improve the image quality of CIS linear array camera, the multi-point one-step correction method is proposed based on gray two-point correction method, using FPGA in the high-speed image acquisition at the same time correcting image, the time delay of correction is much smaller than the computer processing time, it can complete the correction among one driving pulse period only need 0.125 microseconds, and the image uniformity is increased about 10 times after corrected. Compared with the traditional two-point method uniformity is increased by 3 times; the chromatic aberration of CIS camera is basically eliminated.3. Wide-format Multi-segment CIS images combination TechnologyAccording to the growing demand of CIS linear array camera format increasing for automatic optical inspection system, wide CIS camera images combination is researched, image stitching error sources and types are analyzed. Make multi-channel FIFO image mosaic of reconstruction programs and full data acquisition hardware resources, improved the efficiency of multi-channel wide CIS data combination. Wide CIS camera 600 mm format, under 600 dpi optical resolution, USB 2.0 interface transfer rate of up to 30MB/s, use Camera-Link interface, transfer rate of up to 120MB/s. Z-Shaped structure of multiple wide-CIS research on image stitching error characteristics of linear array camera, submit its program and to achieve a fast mosaic image data, submit its program and to achieve a fast mosaic image data, less time consuming, Stitching error is 1 pixel or less and meet the requirements for accurate and fast mosaicking. Wide-format CIS image mosaic effect of technology has been used in actual AOI equipment.4. AOI image detection technologyBased on researching on registration algorithm and color classify of AOI defects image processing algorithms, considering of CIE-Lab color model of the human eye senses, image registration algorithm of regional connectivity is proposed. The features of connectivity are integrated and merged; matching feature points are extracted; the most excellent three pairs of matching points are filtered out combined with the smallest root mean square error method. Bayesian decision theory is applied to the standard model of AOI system and detection extraction. The experimental results show that detection is accuracy and identification is good.Through studying on four areas mentioned above, a number of important issues of AOI projects were solved. This provides a strong reference from theory to support the implementation of the conversion project. Combined with the industry for the next application needs, it accumulates valuable experience of enhancing the level of equipment localization industry.