The Identification And Realization Of Printed Circuit Board Based On Machine Vision Technique

The content of this thesis studies is the measuring which based on computer vision technology circuit board defect of the machine,it mainly realizes with the pattern-recognition method through digital pattern process. The board defect of the circuit measuring includes two parts: The solder joint defect measuring and measuring of the components and parts. The detection procedure can be divided: 1. Formulation and save of the board information of the standard circuit; 2.The board which wait to measure worthy of being accurating and standard board; 3. To draw the board which wait to measure the characteristic information; 4. To compare with the board which measure information and standard board information. Among them, what solder joint information is drawn is its color characteristic information. The characteristic of the component has drawn and adopted color information, form information and projection information. To the localization of the circuit board and the match problem, it has adopted the hole which is on the circuit board to carry on the scheme of the localization. In addition, the component marks and calls discernment that is worth the important part that the component measures, it combine and construct BP neural network and realize marking the claiming value to discern. This defect detection method has low demand of orientating.in clockwise or anti-clockwise, it rotate and will do between 45 degrees, and the characteristic gauge that waiting to measure the board picture of the circuit of different resolution ratios.