Research On Recognition Method Of Abnormal Frequency Point In Electromagnetic Spectrum Of High-density PCB

With the circuit technology coming into the era of the miniature package, large scale integrated and high density assembly, the high density complexity and multilayer of the circuit weakened the physical accessibility of the circuit nodes gradually, and the testability of the circuit and the system fell sharply, and the test costs of board level among the total costs of the electronic circuit kept increasing, and the common test method faced the increasingly serious test difficulty. A new method using the electromagnetic scanning technology for high density electronic circuit fault detection is rising, of which the abnormal frequency identification is the first and key technology of the test method.This paper first studied the features of the scanned electromagnetic spectrum image during the process of testing the circuit, and aiming at the distinguished character that the color of the spectrum curve in the electromagnetic spectrum image of the scanned circuit differ with the background color obviously, we design the HSI color model by using quantitative method to extract the spectrum curve; Then according to the use of spectrum curve in the spectrum diagram and its continuity characteristics, we can obtain the chain codes of the effective discrete points in the spectrum curve of the standard board and the tested board; And then calculate the difference of the effective discrete points between the standard board and the tested board’s spectrum curve, to get the effective discrete points of the spectrum difference curve, and using the effective discrete points of difference to generate spectrum curve; Finally, using niche genetic algorithm, and combined with the characteristics of scanned electromagnetic spectrum image of the high density electronic circuit board to recognize the abnormal frequency of the board.Theoretical analysis and simulation results show that the proposed algorithm is accurate and efficient, and it can automatically detect the abnormal frequency points of the scanned electromagnetic spectrum image of the high density circuit board, and it provides important basis for further use of electromagnetic scanning technology to detect fault components in the circuit.In a follow-up study work, we will improve and perfect the whole process by using the magnetic scanning technology for detecting fault components in the circuit board, and emphasize on the research of the recognition of the abnormal region in the board and the decision problem suspected fault components in the board.