Operation Environmental Perception Of Intelligent Mining Excavators Based On 3D Point Clouds

The mining excavator refers to the excavator that performs the two functions of excavation and loading in the mining operation of the open-pit mine.The comprehensive performance of the mining excavator directly determines the mining efficiency and safety of the entire mine.The excavation operation of the existing mining excavator is completely manually operated.Due to the large size of the mining excavator and the long sight distance of the operator,the operator has obvious insufficient in the ability to perceive the external environment of the excavator,estimate the internal state,and make the excavation decision.The research on intelligent mining excavators is carried out to solve these problems.In the research of intelligent mining excavators,operation environmental perception is the premise and foundation of trajectory optimization,motion control,and autonomous operation.It is the top priority of the research.In this paper,by studying the operation environmental perception system of intelligent mining excavators,the operation environmental perception is realized for intelligent mining excavators.The main work of this paper includes three parts:(1)In order to realize the operation environmental perception of intelligent mining excavators,according to the situation of operation environment and the characteristics of excavator operation,this paper develops a hardware platform to provide the basic hardware support for environmental perception.The hardware platform includes multiple sets of lidars,industrial cameras,inertial measurement units and satellite positioning devices.The temperature of the excavator operation environment is low at night,and the mechanical vibration of the excavator is large.These have a great adverse effect on the work of the hardware equipment.The hardware platform built in this paper has good reliability,strong stability and high precision.The hardware platform can better adapt to various conditions of the operation environment,and continuously provide high-quality and high-precision point clouds,images,attitude and position information when the excavator is operating.(2)In order to cooperate with the hardware platform,this paper develops a software platform to collect and store data of different hardware and realize a variety of functions.The software platform can fuse point clouds and images in real time to generate color point clouds,and can also register point clouds in different coordinate systems scanned by multiple lidars.The software platform also includes the environmental perception function.This function can identify the material piles and trucks in the excavator operation environment,which lays the foundation for the excavator motion control.The display part of the software platform can display the results of data collection and processing in real time,which is convenient for real-time monitoring of the operation of the excavator.(3)In order to improve the recognition accuracy of mining and loading targets in the operation environment,an environmental perception algorithm is designed in this paper.The environmental perception algorithm consists of two parts: multilevel ground extraction and mining and loading target recognition based on feature calculation.The multilevel ground extraction uses a multi-layer structure.The ground can be extracted step by step through local shape analysis,surface point normal segmentation,and surface point position segmentation,which can accurately remove the ground.The mining and loading target recognition based on feature calculation is carried out after the ground extraction.Firstly,the objects in the point cloud after ground extraction are divided into multiple fragments.Then the mining target is extracted.According to the geometric information of the material pile,the point cloud of the material pile is extracted from the fragments.This can help the excavator determine the location of the material pile in the operation environment.Finally,the loading target is recognized.The truck in the fragments after scene segmentation is determined according to the actual truck characteristics.Through the above two parts,the excavator can successfully perceive the location of the operation objects in the operation environment and realize the environmental perception.