Research On Detection Of Rail Target Based On Millimeter Wave Radar And Visual Information Fusion

The construction of intelligent railway is the trend of railway development at present.As a part of the construction of intelligent railway,the safety of freight train shunting and docking has become a topic that is worth paying attention to.Due to the large mass of the train,it still has enormous momentum at low speeds before docking.Therefore,when the train runs near the docking area,the driver need to pay attention to the speed of the train and the distance between the train and the target train to avoid collision and rollover.In addition,other personnel are required to monitor and command the whole docking process,which is relatively inefficient.Therefore,to replace the manual ranging operation and increase the safety of train running process,a real-time sensing and ranging system based on millimeter wave radar and machine visual information fusion for the forward rail target is presented.The system focuses on automatic detection,location and ranging of rail target.This paper mainly includes the following contents.1)A two-step target detection scheme is proposed to detect target train based on the obtained millimeter wave radar information.Firstly,the millimeter wave radar target data preprocessing is completed based on geometric constraints of radar information and the primary target set of millimeter wave radar is received.Secondly,the life cycle algorithm is presented to further filter the target and obtain effective target set.2)For the detection of target train based on the obtained visual image information,an on-rail identification method for trains based on Progressive Probabilistic Hough Transform and density-based spatial clustering algorithm is proposed.And the bottom position of the docking target train is recognized based on the change of rail gray level and is modified based on Kalman filtering.3)A time-space synchronization algorithm is proposed to complete the time fusion based on least common multiple principle and the spatial fusion based on coordinate transformation.Then the joint calibration of the millimeter wave radar and monocular vision camera is completed.On this basis,the detection,location and ranging of the target are realized based on the information fusion of millimeter wave radar and vision.4)An engineering prototype of millimeter wave radar combined with monocular vision camera is developed by using the method proposed in this paper.The experimental results show that the ranging accuracy of the engineering prototype is95.0%.And the ranging algorithm has strong environmental adaptability and robustness in multiple scenes and weather.The above research achievement provide technical support for the realization of unmanned intelligent railway,which has important theoretical significance and practical value to promote the informatization and intelligentization of railway transportation system.