Research On The Multi-source Collaborative Monitoring Method For The Large Deformation Landslide In Fushun West Open-Pit Mine

Presently,the open-pit mines in China have gradually entered the stage of deep mining.Frequent landslides of open-pit slopes have become one of the most serious geological hazards affecting the mine safety production.The Fushun west open-pit mine,mined in 1901,is the largest open-pit mine in Asia with the range about 6.6km long,2.2km wide,and 500m mining depth.The high-intensity mining during more than one hundred years led to the complexity and deterioration of the geology condition and the frequent occurrence of the small scale landslides.Since 2009,the ground fissure appeared in the Qiantai area and expanded continuously,leading to the initial landslide on the south slope.The rate of landslide movement was accelerated obviously after the torrential rain in August,2013.To the end of 2013,the deep-seated rock landslide with the volume of approximately 1×108m3 Was formed and the length of the fissure on the trailing was about 3.1km.The sliding surface drop was more than 18m,and the sliding rate was larger than 100mm/day.The huge landslide seriously threatens the safety of life and property of mine production and residents of surrounding enterprises.The real-time and effective monitoring to this lager deformation landslide is the main problem for the landslide control.Considering the large scale,high sliding rate and the complexity of the geo-structure of the south slope of the west open-pit mine,it is difficult to evaluate the temporal and spatial characteristics accurately according to the single and traditional method of GPS.Although the displacement field at the initial stage of the landslide can be obtained by the D-InSAR method when the displacement value is relatively small.With the increase of the displacement,the problem of phase unwrapping difficulties and phase miscorrelation lead to the error of the displacement field by the D-InSAR.Therefore,on the basis of the National Basic Research 973 Program 'The Air-space-ground Multi-sensor Collaborative Monitoring Method based on Events Driver' of the China Ministry of Science,the comprehensive method including the GPS real-time monitoring system,the high resolution remote sensing image and the ground 3D laser scanning technology was adopted to the landslide monitoring.The combined,synergistic and effective analysis on the ground deformation features of the landslide was performed based on the geology condition of the open-pit mine.Thus,the multi-sources and three-dimensional cooperative monitoring technology mode was explored with the combination of 'points and surface','horizontal and vertical displacement field' and'temporal and spatial analysis' in the open-pit mine landslide with large deformation.The main work are summarized as follows:(1)Study on deformation monitoring of the landslide in the south slope of west open-pit mine based on real-time GPS.The GPS monitoring stations were arranged in the key area on the south slope for the real-time and continuous monitoring.The GPS results could provide the accurate temporal feature of the landslide and the basis for monitoring and early warning of the disasters.The results showed that:there is a temporal feature of‘low rate sliding-accelerate-decelerate' with three stages from April 13,2013 to July 1,2015.The northward sliding rate was in the range of 10-20mm/day in the low rate sliding stage(from April 13 to August 17,2013).After the torrential rain in August 16,2013,the continuous accelerated sliding began,and the average northward sliding rate was larger than 100 mm/day till December 1,2014.The value of the displacement was in the range of 37.5-54.3m in this stage,and the displacement on the slope increased from the east region to the west region with obvious regional feature.In the decelerated sliding stage(from December 2,2014 to July 1,2015),the northward sliding rate decrease to about 20mm/day with the displacement value in the range of 4-6m.(2)Experimental study on landslide deformation field extraction based on the image features matching method.According to the similarity principle,the physical simulation landslide experiment was designed.And the feature points as well as the vector displacement variation information in the optical images at different landslide stages were extracted by the SIFT,CSIFT and ASIFT algorithms.The experimental results showed that:all the three algorithms can be applied for the deformation field extracting at the middle and later stage of the landslide when there is obvious deformation.However,there is lack of corrected points and a higher rate of false matches in the results extracted by CSIFT and SIFT algorithm,respectively.Compared to the results of SIFT and CSIFT algorithm,there are obvious advantages for the ASIFT algorithm in the aspects of the corrected rate and numbers of the matching points,and the displacement accuracy.According to the experimental results,the ASIFT algorithm was the best optimal image feature matching algorithm for the landslide monitoring,which could provide the foundation for the subsequent deformation field monitoring of large deformation landslides based on high resolution remote sensing images.(3)Study on the horizontal deformation field extracting of the landslide based on the high resolution remote sensing images.According to the aforementioned experimental results,the Pleiades high-resolution remote sensing images at different phases in the study area were obtained.And the feature matching algorithms were adopted for the deformation field extracting in the large deformation stage.The results showed that,when the large deformation appeared on the landslide in the continuous accelerated sliding stage,the matching effect of CSIFT algorithm is poor.The number of the accurate matching points extracted by ASIFT algorithm is 4 times than the results calculated by SIFT algorithm.And the accuracy can be improved from 82%to 94%.The feature points matching results further confirmed that the ASIFT algorithm is superior to the other two algorithms in the aspects of displacement extracting accuracy,and the definition of landslide boundary and dangerous areas.According to the results,the deformation field monitoring method based on the high resolution remote sensing images and ASIFT algorithm is proposed for landslide.And this study can provide a new technical approach to extract the surface deformation monitoring in the landslide.(4)Study on the vertical deformation field extracting of the landslide based on the 3D laser scanning technology.The vertical deformation can not be detected by the RS image because of the satellite observation mode.In order to realize the comprehensive monitoring of the south slope landslide,the cloud points of the south slope at two phases were obtained by the three-dimensional laser scanner.The vertical deformation results were extracted by the difference DEM results.The elevation variation information on the slope surface can be visually displayed by the color legends with the elevation accuracy of about 0.1m.The front area and the boundary of the landslide can be accurately delineated by the elevation change statistics method based on the vertical deformation results.(5)The comprehensive analysis of multi-source monitoring data of the south slope landslide.Based on the results of the GPS real-time monitoring,the horizontal deformation field extracted by the high resolution RS images,and the vertical deformation field extracted by the 3D laser scanning,the comprehensive analysis was performed.The results indicated that:firstly,the temporal variation features of the landslide was obtained by the GPS real-time monitoring.Then,the horizontal and vertical deformation field can be extracted by the results of both the RS images and the 3D laser scanning.The spatial range and the boundary delineation of the landslide can be realized efficiently.The temporal-spatial evolution characteristics of west open-pit mine can be revealed by the multi-source results,which could provide the foundation for the landslide early warning and engineering control.(6)Discussion on the collaborative monitoring model for large deformation landslides.According to the research results of south slope landslide in west open-pit mine,the multi-source and three-dimensional collaborative monitoring technology model for large deformation landslide surface deformation is proposed.The temporal variation features of the landslide can be obtained based on the GPS real-time monitoring where the stations were arranged in the key area on the slope.The horizontal and vertical deformation field results can be extracted by the high resolution remote sensing image and 3D laser scanning technology.The spatial deformation field of the landslide can be revealed and the boundary and the dangerous area of the landslide can be delineated based on the combination deformation fields.The temporal and spatial results can provide the foundation for the landslide control.Combined with the‘point-surface','horizontal-vertical',and 'temporal-spatial features',the comprehensive features of the large deformation landslide can be acquired.This study has the reference significance for similar landslide monitoring.