Research On Infrared Image Registration Algorithm For Electronic Board Diagnostic System

Airborne equipment inspection and maintenance is an important link of the civil aviation industry. Timely and accurate for detection of aircraft components is important guarantee of the civil aviation flight safety, and the electronic board maintenance is an important part among them. However, the traditional detection methods due to the low efficiency, high complexity, and the contact detection method will cause certain damage to precision circuit board. This method can't well adapt to the development of civil aviation industry in China.Therefore, only a highly efficient, noncontact detection method can meet the requirements of intelligent, accurate and quick saving in airborne electronic equipment repair. Infrared thermal imaging technique as a method for no contact fault detection method is more and more used in various fields because of its no physical contact, simple operation and intuition.Based on fundamental Infrared Radiation(IR) theory, the infrared diagnosing technology is derived of; In this thesis, the thermal pattern of the circuit board and the diagnosing ability of Infrared Imaging Technology are analyzed; The airborne electronic diagnostic system is designed; In the diagnosis system, the Infrared image registration problem is discussed in depth.In the diagnosis system of airborne electronic failure, image registration is an important part of the image processing and the matching results directly affect the fault location detection of the chips in the next step. In the image registration based on the theory and technology,this paper focuses on the use of the method based on the combination of morphological gradient and mutual information in infrared thermal image registration. This method not only considers all voxel information, and effectively combines pixel position in the spatial space. In the process of parameter optimization, for traditional PSO easy to fall into local extreme disadvantage, this study proposed the combination of the methods about PSO(Particle swarm optimization, PSO) algorithm for global optimization algorithms and the Powell algorithm for the local optimization algorithm.The registration results of the former can provide a very effective initial point to the latter algorithm optimization. The method can not only effectively restrain the local extreme, but also can speed up the convergence. The experimental results show that the method can effectively improve the matching accuracy and speed.