Research On Conveyor Intelligent Monitoring Methods Based On 3D Point Cloud

Real-time and accurate status monitoring of belt conveyors is crucial for the safe and efficient operation of a material transportation system.Belt conveyors are often used in harsh environments,and multiple complex factors pose huge challenges to achieving reliable and accurate monitoring.To get through the bottlenecks such as low accuracy and environmental sensitivity in current methods,the research on conveyor non-contact monitoring is carried out using 3D point cloud processing as the core technology from two aspects: bulk material flow and conveyor belt.Considering the operational characteristics of the conveyor and leveraging the advantages of 3D point cloud data in characterizing the target object’s shape and position,an intelligent monitoring scheme is designed and models are constructed around three key functions:measuring bulk material flow,monitoring the belt’s longitudinal rip,and identifying belt’s deviation.Then,with the aid of extending and integrating them,a unified multiaspect monitoring system is formed.The main research points are as follows:(1)Overall scheme design of conveyor monitoring.The overall monitoring scheme for the belt conveyor is designed.Fisrt,the way to obtain 3D point cloud data and the hardware layout are introduced,and a line-scanning monitoring mode with synchronized up and down view fields is formed.Then,radio frequency identification technology is introduced to mark the conveyor belt,and a unique segmented positioning and recognition mechanism is generated.Moreover,to solve the inaccurate measurement caused by the contact speed sensors,the conveyor belt speed correction modes are designed.(2)Research on measurement for bulk material flow.Considering the dynamic changes of the belt’s surface and the bulk material flow,the measurement method based on 3D point cloud is studied from two aspects: volume and bulk density.First,a volume flow measurement model is constructed using the microelement accumulation method and the space volume difference method.Then,the fluctuation degree of the point cloud is evaluated through the steps of gridding the point cloud data,obtaining the sinkingremove vector,and calculating the characterization value of the fluctuation degree,sequentially.The relationship between the fluctuation degree of the point cloud and the material bulk density is determined through sample fitting.In this way,a dynamic bulk density determination model is established.Finally,the effectiveness of the bulk material flow measurement method is verified using multiple types of bulk material as experimental objects.(3)Research on monitoring for conveyor belt longitudinal rip.To overcome the shortcomings of current non-contact methods such as poor real-time performance,misidentification,and lack of effective quantitative evaluation,a longitudinal rip monitoring method for conveyor belts based on 3D point clouds is studied.First,a frame-by-frame line scanning mode for longitudinal rip’s real-time identification is constructed by longitudinal rip edge suspected points extracting,clustering,container eliminating,and empirical criterion discriminating,orderly.Then,with the aid of principal component analysis of longitudinal rip edge points,the rip is characterized in3 D space to obtain quantization information reflecting its severity.Finally,the effectiveness of the longitudinal rip monitoring method for identifying and characterizing faults is verified through a large number of experiments.(4)Research on monitoring for conveyor belt deviation.To improve the accuracy and reliability of conveyor belt deviation monitoring,a corresponding method based on3 D point cloud is studied.First,the range of deviation monitoring is limited via passthrough filtering.Then,the suspected edge points on the left and right belt sides are extracted by analyzing the distribution rule of belt edge points and according to the slope characteristics of the conveyor belt surface.Subsequently,incorporating the belt’s distance characteristics in the width direction and belt’s slope characteristics,hard and soft segmentation modes for extracting conveyor belt edge points are constructed respectively,which are to obtain the belt’s left and right edge information for real-time identifying deviation faults.Finally,these two different segmentation modes are compared,and the effectiveness of the deviation monitoring method is verified by experiments.(5)Functions’ extension and integration.For the sake of maximizing the effect of the monitoring system,the three major monitoring functions mentioned above are extended and integrated.First,the aforementioned series of models are further extended to form bulk material flow status,belt defect,and belt offset status monitoring models.Then,in accordance with the coupling relationship between bulk material flow and conveyor belt monitoring,the synergy and complementarity mechanism between various functions is studied.Finally,the response mechanism for abnormal status is established and the multi-in-one intelligent monitoring system is formed.The research on conveyor intelligent monitoring methods based on 3D point cloud gives solutions with the advantages of reliability,real-time,accuracy,practicality,economy,and on-site adaptability for the intelligent transformation of belt conveyors.Furthermore,the research also contains a wealth of experiences and skills in applying innovative results to solve practical engineering problems,and it can provide references and inspirations for industrial intelligent development.There are 97 figures,10 tables,and 147 references in this dissertation.