| Mountain geological disasters such as landslides,mudslides and collapses are potential hidden dangers affecting the safety of nearby roads and people.The important agencies responsible for the prevention and control of geological disasters by the relevant departments such as the Bureau of Land and Resources,how to quickly and accurately carry out monitoring surveys according to the characteristics of sudden geological disasters in the region,are important issues that need to be resolved.In view of the shortcomings of long road lines,wide distribution of slopes,and the shortcomings of conventional geological surveys,such as low timeliness and long periods,a deep learning framework was used to train and learn highway landslide disasters,and achieved daily survey functions and rescue functions during disasters.Improving the road safety factor and reducing the unnecessary labor force of the surveyors.According to the problem of the universality of landslides in mountainous areas,the main causes of landslides are analyzed based on the types of landslide-inducing factors and the simple coupling relationship between them.The commonly used monitoring and surveying models and disaster assessment methods for highway landslides are described.Based on the analysis of the types and impacts of typical geological disasters,in order to solve the problem of image blur caused by complex weather problems such as foggy and insufficient illumination in mountainous areas during drone aerial monitoring,the dark channel defogging algorithm and super The resolution reconstruction algorithm is combined and image quality assessment is performed using the no-image reference.Multiple optimization-enhanced experiments are performed on images that are blurred by this factor.The experimental data clearly shows that the algorithm is efficient and convenient.Based on the above,this paper carried out the daily monitoring and emergency rescue functions of the road through the UAV equipped with mobile machine vision,and carried out non-contact monitoring of soil moisture,slope and other landslide indicators on the road around the geological conditions,through the SLIDE model.The stability coefficient was evaluated on the detection area,and the contactless safety assessment of the landslide body was realized.In view of the shortcomings of long road lines,wide distribution of slopes,and the shortcomings of conventional geological surveys,such as low timeliness and long periods,a deep learning framework was used to train and learn highway landslide disasters,and achieved daily survey functions and rescue functions during disasters.Improving the road safety factor and reducing the unnecessary labor force of the surveyors.For the disadvantages of complex coupling factors and large uncertainties caused by geological hazards,the advantage of gray clustering is used,combined with the advantages of information entropy in information theory,this paper adopts a gray clustering method based on entropy weight and interval gray number.Calculate the size of the weight of the hazards and provide the coupling relationship between the harm factors.This method is scientific and effective.This paper proposes and designs an intelligent survey method and system for mountain geological disasters based on UAV-based mobile machine vision and artificial neural network multi-channel information fusion to achieve video conferencing,individual soldier detection,drone detection and identification,and rescue agencies and remote The functions of call work,parameter setting,log storage,etc.,and the experiment of non-contact monitoring of soil moisture and slope gradient by machine vision were carried out to reveal the linear relationship between water content and image characteristics,and provided for the relevant departments such as the Ministry of Land and Resources.Relevant mountain and road geological hazard monitoring survey and information support,to achieve prevention and related emergency assistance and decision-making in the region. |
PDF Full Text Request