**7. References**


**23** 

*China* 

**Applications of Computer Vision** 

Yangjie Wei1, Chengdong Wu2 and Zaili Dong3

*Robotics,Institute of Automation, Chinese Academy of Sciences 2School of Information Science & Engineering, Northeast University 3State Key Laboratory of Robotics, Shenyang Institute of Automation,* 

*1Graduate School of Chinese Academy of Sciences & The State Key Laboratory of* 

Nowadays, micro/nano science and technology has been one of the most attractive research fields. However, real time and accurate observation in micro/nano manipulation is a top important enabling technique. Most recently, with the great development of microscopes and computer vision techniques, real time visualization, including 2D motion measurement

As for 2D motion measurement, visual motion measurement on micro/nano scale is still an open problem. Many researchers have designed different algorithms, and most of them are based on a block matching algorithm, which locates matching blocks in a researched digital image for the purposes of distance or similarity estimation. Usually, block matching based methods can achieve better performances when the texture is not relevant, or the aliasing problem in the derivative estimation, which is caused by the large inter-frame displacements (Giachetti & Torre, 1996). Images, however, are typically processed assuming a uniform grid of pixels. While straightforward, the uniform grid representation does not scale well in a multi-scale setting, because it requires an excessive amount of refinement to capture small details in a image, including sub-pixel resolution. The motion to be estimated is, on most situations in micro/nano manipulation, small and not integer. Therefore, it is necessary to improve the existing algorithms and obtain higher precision not limited by the pixel dimension, i.e., sub-pixel motion estimation. In 1989, Anandan reaches a sub-pixel precision by locally approximating the difference function with a quadratic surface and published (Horn, 1986; Horn & Schunck,1981; Singh, 1990), however, the sub-pixel

As far as 3D reconstruction is concerned, depth measurement, i.e., methods to attain 3D information from 2D images, is an important research field in computer vision, and now it has been one of the key techniques in many fields, such as medicine, robotics, remote sensing and micro/nano manipulation. In recent years, there are various 3D reconstruction methods, including volumetric methods, depth from stereo (DFS), depth from focus (DFF)

and depth from defocus (DFD)( Yin,1999), researched and used in real applications.

and 3D reconstruction, on micro/nano scale is becoming possible.

estimation resolution usually introduces more computational burden.

**1. Introduction**

**in Micro/Nano Observation** 

*Chinese Academy of Sciences,* 

