The Research On The Technology Of PCB Infrared Image Enhancement

With the rapid development of civil aviation transportation industry, the development of aircraft maintenance industry is promoted greatly, and the most common fault is printed circuit board(PCB) on-board. With the development of infrared technology, researchers have placed much importance on the PCB fault diagnosis based on infrared technology. Because the PCB infrared image has disadvantages of strong noise, edge blur and low contrast, the following segmentation, recognition and registration of infrared image are made difficult. So the enhancement of PCB infrared image is a very important part in infrared image processing.The noise, histogram, contrast and resolution ratio of PCB infrared image is discussed,then the common infrared image enhancement method from three aspects of de-noise,improving contrast and sharpening edge is studied, finally the effect of different algorithms by experiments is compared in this paper.On this basis, a comprehensive algorithm based on compressed sensing and the law of universal gravitation is proposed in this paper. Firstly, compressed sensing is used to de-noise.The core idea of compressed sensing is to use the sparse representation of signals to ensure the structure of the original signals and then to reconstruct the original signals accurately.Next, the edge detector based on the law of universal gravitation is taken to implement the edge sharpening. Compressed sensing only has effect on white Gaussian noise, but is not effective on salt and pepper noise; however the edge detection operator based on the law of universal gravitation not only realizes edge sharpening but also has a good effect on salt and pepper noise, making up for the deficiency of the compressed sensing. At last, histogram equalization is applied to improve image contrast.The experiment result shows that through analysis of different PCB infrared images,PCB infrared image is comprehensively enhanced by the method in this paper.