High-speed subpixel image sensing technique and its applications

1.提出了一种基于相位相关（POC）函数的高精度图像匹配方法。新开发的技术可以通过拟合的相关峰的封闭形式的分析模型，以实际的二维数值数据，以提高匹配性能。该技术也可以扩展到频谱加权POC技术，在该技术中，我们使用一些加权函数来修改交叉相位谱以提高配准精度。通过对平移位移、旋转角度和比例因子的估计误差进行评估，对该方法的匹配性能进行了系统的实验分析.研究了基于POC的图像匹配/配准技术在以下方面的应用：（i）电子显微镜的高精度尺度估计;（ii）金属构件检测的高精度测量;（iii）视频序列的超分辨率处理;（iv）高质量背景子画面生成;（v）视频对象分割;（vi）指纹/3D人脸识别.系统实现技术进行了研究，开发上述应用程序。我们研究了应用程序（i）在通用处理器上的实现，因为它只处理2D静态图像。另一方面，我们研究了应用程序（ii）-（vi）在数字信号处理器和特定硬件上的实现。

1. A high-accuracy image matching technique using Phase-Only Correlation (POC) function was developed. The newly developed technique can improve matching performance by fitting the closed-form analytical model of the correlation peak to actual two-dimensional numerical data. This technique can also be extended to a spectrum weighting POC technique, where we modify cross-phase spectrum with some weighting functions to enhance registration accuracy. We provided systematic experimental analysis of the matching performance of the proposed technique by evaluating the errors in estimating the translational displacements, the rotation angle and the scale factor.2. Some applications of the POC-based image matching/registration technique to (i) high-accuracy scale estimation for electron microscopes, (ii) high-accuracy measurement for metal component inspection, (iii) super-resolution processing for video sequences, (iv) high-quality background sprite generation, (v) video object segmentation, and (vi) fingerprint/3D face recognition were studied.3. System implementation techniques were investigated for developing the above applications. We studied the implementation of the application (i) on a general-purpose processor since it handles only 2D still images. On the other hand, we studied the implementations of the applications (ii)-(vi) on a digital signal processor and specific hardware.