The Segmentation Of Infrared Image Of PCB Based On Mathematical Morphology

The accurate positioning of faulty chip is needed in the process of fault diagnosis of PCB. Image segmentation is an important part as the premise of feature extraction and target recognition in the image processing.The densely layout of chips so that the adjacent ones radiate each other in the PCB; The contrast of target and background of infrared image of PCB is low, and the contour of component is not clear; Signal to noise ratio of infrared image SNR of PCB is low. The traditional algorithm is sometimes inaccurate when it is used in segmenting the infrared image. With the integrated density of the electronic equipment increasing, the complexity and difficulty of the fault testing of the circuit board increases. Now, the component level of segmentation for PCB is failed to achieve, the test for PCB can only do for board-level testing. The situation leads to that the original maintenance means and experience can't meet the modern aircraft electronic equipment's maintenance requirements especially in Civil Aviation.In view of the question of the infrared image segmentation of PCB, a new method of infrared image segmentation is proposed with our infrared thermal imaging technology and signal loading technology: The segmentation of infrared image of PCB based on Mathematical Morphology.The paper introduces the principle of the infrared imaging technology and the fault detection principle of PCB, making the research of the feasibility of the infrared diagnostic circuit board; introduces the characters of infrared images and the ImageIR thermal cameral; estimating the background image by Mathematical Morphology, then the original image is substracted by the background image and get the difference image and enhance it; the enhanced image is segmented by Maximum variance between(Otsu) segmentation method to get the binary image;Finally,feature extraction of the object is done by taking advantage of the character of core heating zone of chip to get the target image to be detected.Compared with other several classic image enhancement algorithms, Simulation results show that: the proposed algorithm enhance the brightness and reduce the noise of the image,the segmentation accuracy and the ability to identify electronic components are improved. The segmentation algorithm is a practical and effective infrared image segmentation method.