Our organization works primarily in the field of visual geometry. Carry out research work on the application of various vision technologies for on-orbit spacecraft.
As space technology has been leaping forward, a wide variety of space structures (e.g., space antenna, space telescope, and space solar array) are rapidly advancing towards large-scale, high-precision and deployable space structures. In order to verify the reliability of the structure and monitor its functional performance, all large space structures will undergo a series of rigorous ground-based or space-based tests. PG shows the advantages of non-contact, large-scale, high-precision, multi-point simultaneous measuring, and others. It especially applies to the fast measurement of large structures movement and deformation. As the space camera thermal control system is rapidly advancing, the PG system has been well applied to a series of LSS ground-based or space-based measurement tasks.
Fig.1. Schematic diagram of research direction
Utilizing machine vision techniques for accurate positioning and orientation estimation of objects, robots, or cameras, such as vision-based SLAM (Simultaneous Localization and Mapping), camera calibration, etc.
Employing cameras and optical sensors for high-precision 3D measurements and scanning, such as structured light scanning, time-of-flight (TOF) cameras, etc.
Researching methods to recover the true size and scale information of objects in images or scenes, such as scale-invariant feature transform (SIFT), multi-view geometry, etc.
Analyzing and matching object shapes with high precision, including shape matching, shape reconstruction, shape alignment, etc.
Utilizing machine vision techniques for precise measurement and analysis of surface roughness and texture of objects.
Performing accurate measurement and unwrapping of optical wave phase information, such as phase-shifting interferometry, multi-frequency grating projection measurement, etc.
Techniques for preprocessing, enhancing, filtering, segmenting, and registering images to extract useful information or improve image quality.
Methods for extracting meaningful features from images, such as edges, corners, textures, and representing them as high-dimensional vectors or descriptors.
Approaches for identifying specific objects or entities in images and classifying them into predefined categories, such as face recognition, object detection, etc.
Algorithms for partitioning images into multiple regions or objects to better understand their content and structure, such as semantic segmentation, instance segmentation, etc.
Techniques for real-time tracking of objects of interest in video sequences and predicting their position and motion.
Inferring the three-dimensional structure and geometry of scenes by leveraging image data acquired from multiple viewpoints.
Generating realistic images using machine learning and deep learning techniques, such as image super-resolution, style transfer, image synthesis, etc.
Simultaneous Localization and Mapping by integrating visual perception and motion estimation to achieve real-time mapping and localization.
Research on understanding, analyzing, and reasoning about complex scenes, human actions, objects, etc., such as behavior recognition, scene understanding, etc.