Research Of Phase Unwrapping Algorithm And Its Application In Phase Measurement Profilometry

With the rapid development of science and technology, in the industrial production process, detection of objects surface defect, size and free surface is more and more frequently,and detection accuracy requirements are also rising, classical contact to the object detection method is no longer applicable, in order to solve the problem, the optical 3D measurement technology arises at the historic moment. Optical 3D measurement technology is a kind of non-contact, high-precision projection structure light or natural light surface topography measurement technology, this technology has been widely used in electronics, automobile,machinery processing, textile and other modern industry, such as biomedicine, human detection, automatic processing, online measurement, and so on.Phase measurement is one of the active measurement methods in optical three dimensional measurement technology, which has the advantage of optical 3D measurement technique. However, the phase information collected from the phase measuring profilometry is through inverse trigonometric function is obtained by calculation, the phase value is truncated in [- π, π ] the principal value region, show a discontinuous distribution. Such measurements obtained with the measured object surface morphology distortion. In order to solve this problem we must phase unwrapping operation. Thus, phase unwrapping plays an important role in the whole measurement process. Therefore, the main research content of this paper is the phase unwrapping algorithm.The phase unwrapping algorithm is mainly divided into local(path dependence, path tracking) and global(independent path, minimum norm) method. Local method is a fast and simple calculation of phase unwrapping algorithm, the biggest drawback of this algorithm is choose a suitable integral path is very difficult, and slow computing speed, such as the branch cut method, the quality oriented method; global method is independent of path optimizationalgorithm for solving the package, not the error region(noise, undersampled, holes, etc.) for processing, thus affecting the phase unwrapping accuracy, such as those based on discrete cosine transform method of least square method, iterative method. So far, there is still no phase unwrapping algorithm can solve all the problems.In order to solve the error area and integral path constraints, and to achieve fast unwrapping requirements, this paper puts forward the new method based on integer least squares phase unwrapping, it is a simple and efficient global algorithm. Different from the traditional phase unwrapping method, this method only deals with the gradient of the phase transition part. Therefore, the algorithm is affected by the propagation and the smoothing of the residual error. In both simulation and experiment, respectively, using improved integer least square method and classic cut branch method, fast Fourier transform(FFT), the quality oriented method(QG), minimum cost flow method(MCF) were compared, draw the following conclusion: MCF precision highest FFT operation speed the fastest, and the integer least square method only need and the FFT method is close to the lower computation time,can achieve and MCF method is accuracy. Due to the integer least square method with good performance, such as simple, robustness, high accuracy and high computational efficiency,and does not require user intervention. Therefore, the phase unwrapping algorithm can be used in real-time applications.