Development Of Image Acquisiton System In ACF Particles Automatic Detection

Liquid crystal display(LCD), as a very popular display in modern times, is widely used in various fields. After 21 centuries, the chip on glass(COG) technology can make the LCD lighter and thinner. In the COG technique, anisotropic conductive film(ACF) is one of the most important material. The LCD screen's quality is strongly dependent on the ACF laminated conductive path. In foreign countries, some countries have already developed the ACF particle automatic detection machines, but the cost is too high and the detection rate is common. In our country, most manufacturers still use the traditional detection method, directly observe by metallographic microscope, which is inefficient and requires a lot of labor. Thus, developing a machine with independent intellectual property rights to automatically detect the ACF particles has great significance. And as one of the ACF particles automatic detection machine's basic and core structures, the image acquisition system is very important too.In this dissertation, the hardware configuration, software architecture and modular design of the overall system of ACF particles automatically detect machine's image acquisition system have been analyzed and designed. At last, the test of systematic modules has been put forward. The main contents of this dissertation are as follows:1. The overall framework of system, including hardware structure and software design, has been introduced. The software system has been designed by the “4+1” view method and implemented by module.2. The linear CCD camera acquisition module has been designed, including selecting the line array CCD camera and its accessories(lens, capture cards, light), serial communication protocols between PC and PLC, and the research of auto-focus algorithm.3. The image processing and display module has been designed. The digital graphics programming techniques and the BMP format image display scheme has been introduced. The problem of slow and slack display of large image has been solved by the details display algorithm.4. The design of the auto focus algorithm has been researched. After the comprehensive analysis of the commonly used image definition evaluation function, according to the characteristic of focus image, the absolute variance function has been used to cursorily focus and the Brenner function has been used to precisely focus. This method not only can guarantee the accuracy, but also ensure the speed of overall computing is fast.In this dissertation, the method of using the auto-focus algorithm in linear CCD camera, and applying it in the ACF particle automatic detection has been put forward innovatively. In the process of focusing, the object move perpendicular the linear camera, and take only one picture reflecting the fuzzy- clear – fuzzy phenomenon, to our knowledge, this method is first reported in China. Comparison with foreign machine, the detection speed of this system reaches the international average and the cost is greatly reduced. So, this system has strong market competitiveness and is more suitable for mass production and popularity.