Robot Vision Group

National Laboratory of Pattern Recognition

Institute of Automation, Chinese Academy of Sciences

Vision Applications in Robot Vision Group

Computer vision is an application-driven discipline, a key measure of its success and usefulness is the development of vision algorithms for various social and market needs. To this end, our group carry out many vision applications in cooperation with government and industrial, such as the State Administration of Cultural Heritage, National Romote Sensing Center of China, National Astronomical Observatories, Nokia Research Center, etc.


Obtaining detailed 3D models of ancient chinese architectures from high-resolution images for digital preservation.




Recognizing new stars from celestial spectra images captured by LAMOST.




Obtaining 3D terrain model from a set of aerial images taken by the aircraft or UAV.




Combining multiple aerial photographic images with overlapping fields of view to produce a large panorama view.




Adding computer-generated objects to live videos captured by the mobile phone.



Interacting with virtual 3D environments in a natural and error-free manner.




Providing positioning and monitoring information for mining by reconstruct the iron mine slope.




Convert traditional 2D films to 3D ones.



Simultaneous localization and mapping with single camera.



Using MicroSoft Kinect to achieve real-time indoor 3D reconstruction.