The Research And Realization Of Circuit Board Image Recognition System Based On Virtual Instruments

This thesis has done some research on LED/LCD circuit board image processing and recognition theory and methods based on virtual instruments, with the background of circuit board testing in factory. With the processing and character feature extraction of the different circuit board image, including the character, numbers and pictures, we use the different methods to recognize the image. This research and application is very useful for the actual LED/LCD circuit board image recognition. It could release the people's job intensity and improve the stability of the circuit board image recognition. This thesis mainly does the research work as follows:I. For the whole research of the circuit board image, we should build a hardware system to acquire the circuit board image and ensure a high quality image. So the thesis introduces the design of the circuit board image recognition hardware system, including the camera selection, background light design and relay circuit board design. And then realizing the image acquisition via the system based on virtual instrument.II. After acquiring circuit board image, the thesis has done some processing on the original image, such as image grayscale conversion, image binarization, edge detection, median filter and image segmentation. This step is preparing for the image recognition with a high quality image.III. For the LED circuit board image, the thesis first does pretreatment on the image, and then do research on the image feature location, image segmentation and number recognition.IV. For the LCD circuit board image, the thesis also does some pretreatment on the image, and finds the valid way to eliminate the influence of the background. Then the thesis uses the improved template matching method to recognize the LCD circuit board image.V. The thesis analyzes the different recognition ways and proposes the improvement measures based on the experiment results. So the wrong recognition rate could be lower down through these improvement measures.VI. The two types of circuit board images recognition software system are both developed on LabVIEW and IMAQ Vision. The software system has implemented many works, such as image acquisition, image processing, image segmentation and image recognition.