该文提出一种Integral Imaging和模拟退火相结合的深度测量方法。针对传统立体视深度测量的复杂性,采用全真三维显示技术Integral Imaging成像技术记录空间三维信息。利用带约束的特征匹配算法并综合多种信息分析Integral imaging视图对之间的视差。将寻找匹配的问题设计为全局能量函数模型,并用模拟退火算法使该能量函数最小化,从而实现所有视图对中所有特征点的匹配。最后结合Integral Imaging深度估算公式获得空间深度信息。仿真结果表明该方法的有效性。
A depth estimation method of combined lntegral Imaging and simulated annealing algorithm was proposed. Aiming at the complexity of the traditional stereo vision depth extraction method, lntegral Imaging was used to record the scene 3D information. Feature matching with constraints condition and integrate multi-information algorithm was applied to analyze the disparity of view pairs extracted from Integral Imaging. Finding corresponding problem was summed up as the global energy function. Simulated Annealing Algorithm was used to minimize this energy function. Then all feature points in one view image were matched with the ones in another view image. The disparity gained by the above method was applied into the depth estimation formula of Integral Imaging. A good result was gained by the proposed method proposed by the paper. The experiment has approved the validity and feasibility of the method.