Automatic Optical Detector Image Acquisition And Processing System

With the rapid development of electronics industry, the printed circuit board (PCB) industry has also gotten speedy technical improvements. The rapid development of PCB products is facing the following aspects: ultra-thin, small components, high-density, and fine pitch, meanwhile the advantages such as high-tech and high complexity have reached a very high level. Therefore, the requirements for the quality of PCB productions have become more and more stringent, and quality control has to be done. Currently, there are various online testing equipmens for PCB production line, some of which are using optical method, others are using electronical method(that is two major categories: optical inspection and electrical inspection). Automatic optical inspection equipments (AOI) using cameras, scanners and so on, scan PCB board to compare the images between standard PCB board and PCB board under testing, in this way, location between holes, hole diameter, circuit width, circuit clearance, defects of electronic components and etc on the PCB board can be detected.In this dissertation an image acquisition and processing system of automatic optical inspection equipment is introduced, which is applied on AOI for PCB images acquisition and processing. And on this basis, this dissertation also makes some depth research on complex image processing algorithms which is applied on FPGA, comes up with a viable program, and provides the corresponding design and simulation.This system uses Contact Image Sensor (CIS) for image acquisition, Which is composed of 3 CIS components length 21 cm, so the total checking width could reach 60 cm. Xilinx Spartan IIIAN FPGA is in charge of controlling the data stream, prepocessing the image after storing data into storage systems, eventually passing the processed image into PC through camera link, and the complex image defects inspection is run by PC.In this dissertation, the implementation process of the hardware system and the programming of the FPGA design are presented. In aspect of hardware circuit: the components have been selected, based on the whole design structure of the system, furthermore the type of chips has been chosen considering the specific requirement; the hardware circuit schematics are consisted of constructing the circuit system, designing power supply and interface for all parts of the circuit, and designing the peripheral circuit for FPGA; considering the theory of transmission line has been introduced, and the signal integrity problem encountered in the high speed circuit design has been analyzed theoretically, including signal reflection, crosstalk, and ground bounce. Six-layers PCB board has been designed considering the theories mentioned above. In aspect of FPGA designing: Xilinx ISE developing environment has been used, and Verilog HDL has been chosen as the RTL synthesizable HDL description language.Aiming to improve the speed of image preprocessing, this dissertation shows us a further research on implementation of complicated image preprocessing using FPGA. Median filtering and sobel algorithm, canny algorithm edge detection image preprocessing circuits are designed using Verilog HDL, and the general simulation is done successfully.The image acquisition and processing system of automatic optical inspection equipment as presented in this dissertation, is applied to detect diameter and amount of the holes for PCB, including image acquisition function and simple image processing function. The circuit board being detected has speed of 5 m/min, length of 60 cm. The system has the following advantages: simple structure, small size, low cost, and powerful function of handling image acquisition and preprocessing.