A new 2D motion measurement method based on neighbor principal feature matching

Traditional 2D motion measurement based on image registration estimates displacement through matching the brightness of each image block, thus causing its matching accuracy and stability on micro/nano scale to be lower. In order to improve the accuracy and stability of 2D motion measurement on micro/nano scale, a new sub-pixel displacement measurement method based on neighbor principal feature matching is proposed, considering small differences in the characteristics of the neighboring image blocks. First, the basic principle of Principal Component Analysis (PCA) is introduced based on analysis on the problems of the traditional image registration, and then neighbor principal feature matching (NPFM) combined with an off-line training is proposed in micro/nano motion measurement. Subsequently, a sequence of simulation is conduced, and the result shows its improvement, compared to traditional image registration in 2D motion measurement. Finally, a high-precision nano platform, a high power microscope and a standard grid are used together to validate our algorithm, and the result shows that the accuracy of the algorithm improves nearly 10 times more than the conventional blocks matching method. What's more, the algorithm has greater robustness in selecting the image block's position and size.