A monocular vision localization method based on unmanned underground mining vehicle using barcode

A method to compute the lean stance of the camera is proposed, which is based on the parallel perspective mapping model of the camera. This method improved the common monocular vision-based localization method and can be used in the operating system of unmanned underground mining vehicle to detect the travelling status of vehicle in real-time, fit the bad underground road condition, and reduce the distance detection error which occurred when the camera is lean to the road. This method computes the lean angle of the camera according to the high of landmark and the interior parameters of the camera, so we can detect the distance between the vehicle and the landmark and reduce the error which occurred if the optical axis. Experiments show that this method can faster detect the lean angle which is one of the camera exterior parameters. It is of high precision and fit real-time processing. This method is easy and effective, can reduce the detection error and is suitable for the requirement of underground mining work.