Research On 3D Positioning Method Of Tower Crane Hooks For Smart Construction Sites

With the rapid development of the smart industry centered around smart cities,the number of smart construction sites is increasing year by year.Construction sites are rapidly developing towards intelligence and convenience.Among them,Tower crane is a common large construction equipment in the construction site.Because of the danger of high-altitude operation,the positioning of tower crane hook is an important task.This article focuses on the smart construction site and takes the tower crane safety monitoring system as the research object.It mainly focuses on the problems of difficulty in obtaining the position of the tower crane hook and potential safety hazards in hook monitoring.The principle research on object detection,point set matching,and visual distance measurement has been carried out,and the threedimensional positioning of the tower crane hook in the engineering site has been completed.The main research content is as follows:Firstly,this article introduces the tower crane safety monitoring system in smart construction sites,mainly focusing on detailed research on hook visualization.Analyze the key technologies related to target detection principles,and on this basis,conduct a detailed exploration of the binocular stereo vision of the tower crane system.The imaging principle of binocular cameras involves four coordinate systems.This article provides a detailed derivation of the conversion relationship between visual coordinate systems,providing reliable theoretical support for subsequent image processing and analysis.Secondly,this paper proposes an innovative solution to the problem of high mismatch rate and poor reliability of the traditional point set matching algorithm based on the Gaussian Mixture model.The local sensitive hash algorithm and piecewise matching strategy are introduced,and the mixing coefficient in the Gaussian Mixture model is improved by combining the characteristic information of the descriptor,which improves the reliability and accuracy of the matching.In order to verify the registration effect of the improved algorithm,the improved algorithm is compared with the Coherent Point Drift(CPD)algorithm and the Non uniform Gaussian Mixture model(NGMM)algorithm.The experiment shows that the improved algorithm in this paper has good registration accuracy and robustness.Then,an improved visual distance measurement strategy for lifting hooks was proposed to address the difficulties in positioning the hooks of engineering tower cranes.Build a binocular vision system,calibrate the binocular camera,and calculate the internal and external parameters and distortion parameters of the camera.Using YOLOv3 network to detect hook targets.And point set matching was performed on the hook target in the image,and the target parallax value of the binocular image was calculated using the corresponding relationship of the feature point set,achieving depth ranging of the target.This provides a solid foundation for the practical deployment of subsequent algorithms.Finally,a three-dimensional positioning strategy for the hook is proposed and algorithm deployment is carried out to verify the technical feasibility of the visual positioning method for the hook proposed in this paper through experiments.Transplant the algorithm in this article on the mobile end and deploy a tower crane safety monitoring system on the construction site.Analyze the results of the three-dimensional positioning scheme proposed in this article and propose further improvement directions.The application of this research provides certain development ideas for the development of smart construction sites.