Research On Slope Surface Displacement Monitoring Method Based On Binocular Vision

Our country has a vast territory and complex terrain.During recent years,some extreme meteorological disasters and people’s unreasonable development of the geographical environment have caused frequent geological disasters.Among the geological disasters,sloperelated disasters have the largest number and the largest impact range.Frequent slope disasters have greatly affected my country’s economic construction and the safety of people’s lives and properties.In order to know the overall state of the slope in time,so as to prevent slope disasters and ensure the safety of people’s production and life,it is necessary to use scientific methods to carry out long-term monitoring of the slope.The traditional slope monitoring method is to collect various data of the slope manually by using various monitoring instruments,which is low in accuracy and time-consuming and labor-intensive.In recent years,the emerging automatic monitoring technologies for slopes mainly include GNSS technology,3D laser scanning technology,synthetic aperture interferometric radar,measuring robot and so on.These methods have their own advantages in the field of slope monitoring,but there are also disadvantages such as difficult equipment installation and high cost.Therefore,in order to develop a low-cost,high-precision,and simple-to-lay slope displacement monitoring system,this paper proposes a new method for slope displacement monitoring based on binocular vision.The main research work is as follows:1.A method for monitoring slope displacement based on binocular vision is proposed.The basic principle of this method is to install several artificial markers on the slope to be measured,and collect images through binocular cameras,and then carry out identification,positioning,and three-dimensional coordinate calculation to obtain the displacement of the marker point.An artificial marker suitable for long-distance monitoring was designed,and the principle of the system layout and the method of judging the slope state according to the monitoring results were introduced.2.The camera is calibrated by Zhang’s calibration method,and the three-dimensional coordinates are solved by using the mathematical principle of binocular vision.The binocular camera is composed of two industrial cameras,and the binocular camera calibration experiment is carried out by using different specifications of the calibration board.According to the experimental results,the factors affecting the calibration accuracy are analyzed,and suggestions for improving the calibration accuracy of the binocular camera are given.3.Using deep learning algorithm and image processing technology to realize the identification and localization of markers in complex background.Images of markers in different environments were collected,datasets were created,and the model was trained by the YOLOv5 algorithm.The trained model could well identify and extract marker images in various complex backgrounds.The positioning of the markers is realized by means of gray histogram averaging,homomorphic filtering,adaptive threshold binarization,Canny edge detection and ellipse fitting.The positioning accuracy experiments of different marker image sizes and the stability experiments of the identification and positioning of markers under illumination changes are carried out.The experimental results show that the identification and positioning algorithm has good accuracy and stability.4.Based on the above research work,a set of binocular visual slope displacement monitoring system was developed,and a series of simulated slope experiments were carried out.Dynamic displacement monitoring experiments were carried out on markers at different distances and under different lighting conditions in the laboratory,respectively,and the monitoring results of the distance and lighting experiments were compared with the stepper motor data;in order to explore the stability of the monitoring system,different multiple markers of distance and location were monitored continuously for 24 hours.The experimental results show that the displacement monitoring accuracy of the system is as high as millimeter level within 20 m,and it has good stability under different lighting conditions,which can meet the needs of actual slope displacement monitoring.