基于同步定位与地图构建的色彩追踪智能小车系统

Research on Color Tracking Intelligent Car System Based on SLAM

采用RGB-D相机,基于ORB特征构建SLAM,用于智能小车的定位与建图,将RGB-D相机提供的彩色图像与表征像素距离的深度图作为主要输入信息,将里程计数据进行融合滤波,得到更为精确的运动位姿作为辅助,在仿真中进行定位导航。以物体的颜色和形状作为基本特征进行识别,通过相机获得的图像来实时动态地识别具有特定简单色彩的物体,使用颜色阈值分割与目标轮廓提取检测相结合的算法和滤波器对结果进行估计。最后通过路径规划实现智能小车的自主导航。

Basing on ORB features and RGB-D camera to construct SLAM for the positioning and mapping of intelligent car was implemented,in which,having color images provided by the RGB-D camera and depth map which representing the pixel distance taken as the main input information,and having the data of odometer and IMU fused and filtered to obtain more accurate motion pose taken as the auxiliary information for location navigation in simulation;in addition,having the color and shape of objects recognized as basic features,and the objects with specific simple colors dynamically recognized in real time through the images obtained by the camera;and having the filter and the algorithm which combining color threshold segmentation with target contour extraction detection used to estimate the results and finally,through the path planning,the autonomous navigation of the car can be realized.