Key Techniques Of Phase Retrieval For ESPI And Experimental Research On Dynamic Thermal Deformation Of Circuit System

Electronic Electronic speckle pattern interferometry (ESPI) is a modern opticalmeasurement technique, which is based on laser technology and combined withcomputer image processing technology. It has the feature of whole-field, non-contactand high resolution, and has been widely used in new materials research, mechanicalproperties test of electronic components and engineering components, componentvibration test and nondestructive testing.With the rapid development of electronic products, various electronic circuitshave been widely used. In the application of electronic products, thermal deformationcaused by the heat generated by the electronic components threatens the reliability ofthe whole circuit system; therefore, the study of thermal reliability of the circuitsystem is particularly necessary and is of a great significance in engineering,especially for real time thermal deformation under working.This dissertation is divided into two parts: firstly the fringe skeleton method inESPI was analyzed and studied, and secondly the experimental research on thermaldeformation of electronic circuit was conducted using dynamic ESPI.The works inthis dissertation are as follows:(1) ESPI fringe patterns contain heavy speckle noise, which degrades the qualityof the image, making the phase retrieval difficult; therefore, exploring the effectiveand fast ESPI fringe pattern filtering methods is one of the key techniques in phaseretrieval. Several classical ESPI fringe pattern filtering methods including Fouriertransform method, Fourier transform method based on fringe orientation, windowedFourier transform were compared and analyzed. Further, by using frequently usedimage quality assessment indexes, quantity evaluation of the filtered results obtainedwith the methods was performed, summarizing their characteristic and shortcomings.(2) In order to improve the efficiency of ESPI filtering method, Curvelettransform method is applied in the application of fringe pattern filtering. By utilizingthe correlation of adjacent scales of Curvelet, combining the advantages of the softand hard thresholding methods, a novel Curvelet transform method with moreadaptability is proposed. The new method was employed in fringe pattern filtering,and was compared with other classical methods. (3) Fringe skeleton interpolation is the most straightforward phase retrievalmethod in ESPI. However, for the phase retrieval from high density noisy fringepatterns, extracting skeleton from those fringe patterns efficiently is another keytechnique in ESPI. The oriented coupled partial differential equation to calculategradient vector flow field is proposed for this purpose. Based on the topologicalcharacteristic of the computed gradient vector flow field, the skeleton from highdensity noisy fringe patterns can be obtained by the method.(4) Experimental study on thermal deformation of electric circuit with dynamicESPI was conducted. In experiments, the thermal mechanical problems in electricalsystem were investigated from different perspectives including temperature-rise andcooling process, different fixed forms, different input voltages, and differentencapsulation modes.