Matlab投影仿真及三维曲面重构实现及演示程序

 这个工具箱非常全面地实现了投影仿真及三维曲面重构。不仅包括算法源代码，还包括演示程序，非常推荐！

曲面重构(亦称曲面建模)是实现反向工程的重要环节"通过建模可以将离散的测量数据重构出连续变化的曲面!

曲面拟合包括曲面插值和曲面逼近!曲面插值要求插值后的曲面严格地通过数据点#曲面逼近是基于离散数据构造出一个曲面"使这些点在某种意义下最为接近所构造的曲面!目前曲面建模领域主要分四边域网格曲面建模和三角域曲面建模!四边域网格曲面建模主要面向有序数据点的曲面建模"而三角域曲面建模是面向散乱数据点的曲面建模!

Photometric Stereo with cast shadows emulation and 3D surface reconstruction

This is a tiny toolbox on Photometric Stereo initial implementations in Matlab, with extention, that including some algorithm of 3D surface reconstruction, based on analyses of distributions of the cast shadows (called a geometric approach to the problem, in opposition to radiometric one, which is obviously P.Stereo methods).

The user, up to some degree, is free to choose its own scenario of data processing within frameworks of patchlet/pointwise Photometric Stereo methods, which are called on later the radiometric approache to the problem of monoscopic 3D surface reconstruction.

This tiny toolbox has the desired feature of rendering with option of the cast shadows (performing even in older versions of Matlab releases).

Finally the four first examples consider only built-in random surfaces, while the next exemplary given scenarios also include the generation of virtual cutting surfaces of abrasive tools.

And what is perhaps the most peculiar feature of this toolbox, it also includes some much evolved random surface generation process, with adoptions of Markov Chains and enourmously extended matrices of transient states probabilities.

Those matrices indicate constructions of directed graphs, which in turn, are altogether built in such manner, as to be some analogy (quite purely and hipothetically) to structures known from nature. For instance, these are directed graphs resembling intermediately the structures of cyclic and policyclic organic compounds.

Consequently, in the result of running of extra added m-scripts, one actually obtains some randomly generated anizotropic surfaces with grooves. The surfaces, where perhaps ductile materials are being plowed.

Apart from ten exemplary issued scenarios (contained within ten m-scripts from ex01st...m to ex10th..m files) the user is free to use self-sustainable m-scripts for rendering virtual images, as well as, for determinaion of vector gradient field, with subsequently occurred global 3D surface reconstruction, please see: 'PatchletPStereo.m' m-script.

RESUMING: this tiny toolbox on Photometric Stereo implementations in Matlab environment lacks only one feature. I mean, it has not thresholding mechanism for the input data of 2D intensity images.