为突破传感源单一、平台不便和设备昂贵的缺陷，并解决同步定位与地图构建中智能化与精确度不可兼顾的矛盾，开展多源（Kinect和DSLR）传感器协同多基线多特征（点线）的室内无人车同步定位与地图构建研究。将从以下方面探索彩色相机与深度相机协同多基线多特征几何遥感反演室内场景三维模型的机理与方法：1) 通过离线相机检校与传感器定向，转化数据为信息、挖掘信息为知识，以构建先验知识库引导后续定位与测图处理；2）运用广义点摄影测量理论建立多基线协同多特征（可视的几何点、不可视的灭点和直线）位姿估计的理论与方法，提升定位与定姿精度；3）基于“辩证法”研究多基线条件下点线特征互为协同匹配的系统方法，增强三维地图构建的自动化和精细度；4）丰富同步定位与地图构建的不确定性与质量控制原理，从局部到整体洞悉不确定性并协同控制系统质量。本课题可促进多学科交叉融合与创新，开辟智能化、精确化同步定位与地图构建的新思路。

In an attempt to break through the limitations of single sensor source, inconvenient platform, the expense of equipment, and to solve the contradiction between the intelligence and accuracy of simultaneous localization and mapping (SLAM), this project will investigate accurate SLAM by synergizing multi-source (Kinect and DSLR) sensors, multiple baseline and multiple features (point and line) for indoor unmanned vehicle. Furthermore, synergetic inversion mechanism and methods of geometric remote sensing based on color camera and depth camera, multiple baseline and multiple features will be explored as below: a) create a priori knowledge base involving offline camera calibration and sensor orientation, which helps to guide subsequent processing by developing a set of data mining methods that will transform data into information and information into knowledge; b) establish theory and methods of a Multi-baseline-and-multi-feature-based position and pose estimation by utilizing generalized point photogrammetry that includes the physical points, the feature line points, and the invisible points into an adjustment model, which can continuously improve the camera’s accuracy of position and pose; c) study dialectics-based systematic method of point-line synergistic image matching, which can enhance the automation and fineness of 3D mapping; d) enrich the principles of uncertainty and synergetic system control by gaining an insight into uncertainty and synergetic controlling system quality from local to global SLAM. The proposed project will promote interdisciplinary integration, innovation, and provide new ideas for SLAM.