Research On The Method Of Measuring Distance Of Tower Crane Obstacle Based On Binocular Vision

Tower cranes are commonly used in the construction site construction equipment,due to the complex working environment,prone to collision,in the event of an accident will often result in serious economic losses and casualties.In the construction process,how to avoid the collision between the tower crane or between the tower crane and obstacles has become one of the hot issues that people pay close attention to.Aiming at the problem of distance measurement in the anti collision of the tower crane monitoring system,this thesis proposes a method of distance measurement for tower crane obstacles based on binocular vision,which has certain practical value and application prospect.In this thesis,we mainly carry out theoretical and practical research on camera calibration,feature extraction and matching and distance measurement system in binocular vision system.First of all,the binocular vision measurement system is established and a binocular vision measurement system is constructed.Then considering the application environment,cost and calibration accuracy,after analyzing the merits and demerits of various camera calibration methods,a binocular camera calibration method based on OpenCV and essential matrix is proposed.The calibration of camera intrinsic parameters is calibrated online by OpenCV programming,and the external parameters are calibrated by matching the feature points of the images,combined with the epipolar constraint equation,the essential matrix is solved in a linear way.This method is fast and simple,and is suitable for the calibration of outdoor situations.And then put forward the SIFT algorithm to extract the image feature points and matching.In view of the shortcomings of the algorithm itself,such as The descriptor dimension is high,the computation is large,and the mismatch rate is relatively high,the PCA and RANSAC algorithms are proposed to improve it.PCA algorithm is used to reduce the dimension of feature points descriptors,and then BBF search algorithm is used to do nearest neighbor search for low dimensional feature vectors of feature points.Then RANSAC algorithm is introduced to remove mismatch pairs.The improved algorithm is simpler and the image matching efficiency is guaranteed.Finally,combined with the above research results,a binocular distance measurement simulation experiment is carried out.And came to Fushun Yongmao Construction Machinery Co Ltd on the tower crane field experiment.The experimental results verify the feasibility of the proposed method and have a certain application value on the problem of anti collision of the tower crane.