A 3D Point Cloud Spatial Localization Method Based on YOLOv5 Dual Camera
With the development of positioning technology,high-precision,low-cost,real-time positioning has become a hot topic in current research.Based on this,this paper proposes a three-dimensional point cloud spatial localization method for du-al-camera objects.In this method,YOLOv5 is used to identify the same object under dual cameras,and then the pixel coordinates of the object are obtained respectively.Then the camera conversion model is used to convert the pixel coordinates into world coordi-nates,and the least square method is used to solve the three-dimensional coordinates to achieve spatial positioning of the object.Fi-nally,Colmap is used to establish a 3D point cloud model to achieve 3D visual positioning of objects in the 3D point cloud and im-prove people's perception of position.The experimental results show that the positioning accuracy error of the algorithm is less than 7 cm in the X-axis motion direction,less than 10 cm in the Y-axis motion direction and less than 16 cm in the Z-axis motion direc-tion.Good spatial positioning effect is obtained.
spatial positioningtarget recognitionleast square method3D point cloud
国家科技期刊平台
NSTL
万方数据
