Research On Visual Perception And Trajectory Planning For Independent Slope Repair Work Of Excavator

Excavator slope repair is a common earthwork operation,the purpose of which is to use the operation of an excavator to change the terrain to meet design requirements.It is commonly used in engineering fields such as road construction,building construction,and water conservancy projects.Currently,the slope repair method is generally carried out by manual operation of the excavator.However,manual operation of the excavator for slope repair still has many problems: firstly,during the slope repair process,the slope and height of the slope to be constructed must be precisely controlled,and any small error may lead to unstable or unqualified slope repair.Relying on the visual measurement method of the driver is difficult to ensure the accuracy of the slope repair.Secondly,the construction site environment is relatively harsh,with dust,noise,long-term exposure to light,and high working intensity posing many health risks to the driver.Finally,there are still limitations in some slope repair work sites,such as environments that are not suitable for human entry and emergency rescue and relief applications.Based on the deficiencies of manual operation,achieving excavator autonomous slope repair is one of the urgent problems to be solved in the current field of construction machinery,which will also promote excavators to develop and innovate towards intelligent and autonomous construction operations,departing from traditional manual operations.This article is about the research on information perception and trajectory planning for excavator slope repair operations.The main content of each chapter is as follows:A method for building a binocular vision information perception system based on autonomous slope repair operations is proposed.Firstly,the overall construction process of excavator autonomous slope repair is analyzed,and the operation process of autonomous slope repair is determined.On this basis,the imaging model and ranging principle of binocular camera are analyzed.Secondly,the influencing factors of slope repair environment are determined,and the AD-Census algorithm is optimized based on this as the basis for stereo matching of the binocular vision system in the autonomous slope repair process.Finally,the designed information perception system in this chapter is experimentally verified.A positioning and fitting method for the excavation line of the slope to be repaired is proposed.Firstly,using the established binocular vision system,images are collected and depth detection is carried out for the sample groove manually built and the slope area to be repaired.Three-dimensional reconstruction is performed by combining their RGB images and depth images,and the standard slope surface,upper construction line,and lower excavation line information of the sample groove are extracted.Secondly,point cloud segmentation technology is used to extract the point cloud of the excavation line under the slope,and a spatial line is fitted as the position for the excavator bucket teeth to stop repairing the slope.Finally,adjustments and optimizations are made to the three-dimensional reconstruction,point cloud segmentation,and line fitting techniques for the harsh environment of typical construction sites,and experimental verification is conducted.A positioning and fitting method for the construction line on the slope to be repaired is proposed.Firstly,images of the slope area to be repaired are collected,and targeted image preprocessing is carried out for typical harsh slope repair environments.Secondly,edge detection is performed on the slope image,and Hough line detection and least squares line fitting are performed based on the edge data.Based on the standard upper construction line fitted by the spatial line in the sample groove,and combined with the position of the lower excavation line and the required slope angle for construction,the position of the upper construction line of the slope is determined.Finally,experiments are designed and the proposed method for detecting the upper construction line in this chapter is validated.A system for real-time monitoring of slope repair and excavator working trajectory planning has been built.Firstly,by comparing the distance information between the standard slope surface of the sample groove and the slope surface to be repaired,using the plane fitted by the standard slope surface as the reference plane,the Alpha Shape algorithm is used to reconstruct the surface of the slope point cloud data,estimate the total amount of earthwork in the slope repair project,and obtain its volume information as a reference.Secondly,a solution is proposed for special soil slope repair methods,and the volume of earthwork already excavated is monitored and fed back in real-time during the slope repair process.The smoothness of the already repaired area is also checked in real-time.Finally,experiments are designed to validate the system and the working trajectory of the excavator during typical slope repair movements is analyzed.