Study Of Shape Measurement By Digital Speckle Temporal Sequence Correlation And Its Application In 3D Shape

A new method of three-dimensional shape measurement called digital speckletemporal sequence correlation is presented. Its basic principle, four methods ofenhancing the precision and quick data-dealing algorithm are introduced in detail.The dominating factors of influencing the precision are analyzed by computerstimulation. At last, the applications of our method to typical shapes are explored.The difficulties encountered when measuring the jumping and space isolated shapeby traditional methods are surmounted and the these shapes can be measured easily,in addition to other complicated shape can also be measured successfully by ourmethod.In chapter one, the traditional 3D sensing methods are introduced simply and theiradvantages and deflects are formulated; in chapter two, the main achievements andthe deflects of digital speckle method are introduced, and the key contents of thispaper are given.In chapter three, the basic principle of digital speckle temporal sequence correlation,speckle-producing method, correlation algorithm and measure operation areintroduced in detail.In chapter four, the dominating factors of influencing the precision, including systemparameters, speckle quality, calculate parameters, surface property, are analyzed indetail by stimulation.In chapter five, four methods of improving precision are presented. The first one,Correlation Weight Interpolation bases on the characteristics of the correspondencerelations between the correlation values and errors distribution; the second one,Multiple Reference-Speckle-Set bases on the principle that spatial random propertycan be transformed into temporal random property; the third one, Multiple PeakOptimization based on the fact that generally there is a correlation peak located onthe position of the true value, and the last one, Local Medium Value Interpolationmainly bases on the fact that object shape can not change rapidly. The further detailsof these methods are described in the paper. In chapter six, factors of influencing datadealing speed are discussed and a quick algorithm to decrease correlation quantitynamed as Local Search is presented.In chapter seven, the measure results of the boss and the dental model are given. Theformer result shows that our method can measure jumping surface easily andprecisely, the later shows that our method can also be used to measure complicatedsurface. From the result of dental model, we can see that our four precisionenhancing methods are very effective.In chapter eight, the application fields and research status of human body 3D sensingare generalized and our method is used to scan human body model in the BMS andthe valuable result is given.At last, conclusions are given, and suggestion about the further work of this topic isalso given.