Design And Development Of Foundation Ditch System Based On Visual Measurement

In recent years,deep foundation pit engineering has been greatly developed in both underground space and high-rise buildings.Due to its large scale and deep depth,deep foundation pits are also areas where engineering accidents are frequent.The frequent occurrence of accidents has raised many technical problems for the engineering industry.In addition to reasonable selection of supporting structures and reliable construction schemes,it is necessary to monitor and warn the deformation of foundation pits.In traditional foundation pit monitoring,total station and theodolite are usually used to monitor the horizontal displacement of foundation pit by using small angle method,or polar coordinate and improved polar coordinate method.However,due to the general requirements of total station and theodolite,the layout of the monitoring points is limited,which affects the accuracy and timeliness of the foundation pit monitoring.From the perspective of visual measurement,this dissertation proposes a foundation pit monitoring scheme based on visual measurement,which provides a new method for the foundation pit safety monitoring program.The research emphasis of this dissertation is divided into following aspects:1.The background and significance of the research are expounded.The status quo of machine vision measurement is introduced from two aspects of software and hardware.The current research status of foundation pit monitoring technology based on machine vision is analyzed.2.From the angle of practical engineering application,the current mainstream of research direction of foundation pit monitoring is analyzed.By comparing and analyzing the advantages and disadvantages of different methods,the advantages of foundation pit monitoring based on visual measurement are summarized.3.The composition of machine vision system is introduced in detail.The core component of the system,namely the data processing system,is optimized.An image processing method of multi-objective optimizing target center is proposed,and by using multi-dimensional feature error compensation method,the error is reduced and the central point is optimized.4.The working principle of the system is illustrated.The deformation is obtained by comparing,analyzing,and calculating the processed image information.By using based BP algorithm fuzzy neural network,the deformation of foundation pit is predicted and forewarned.5.The system is applied in actual projects.First,the deformation data calculated by the system is compared with the monitoring data of the third-party foundation pit monitoring unit to determine the accuracy of the system.Second,the simulation and early warning parts of the system are verified.By comparing the predicting data with the actual monitoring data,the reliability of the simulation system is determined.6.The dissertation summarizes the research content,points out the existing problems in the engineering practice of the system and proposes the direction for further research.