Research On Bridge Crack Detection And Evaluation Technology Based On Real-time Location And Map Construction

UAV technology has been rapid development in recent years.Because of its low cost and flexibility,UAVs are increasingly being used in.areas such as agricultural irrigation,disaster relief,counter-terrorism detection,unmanned surveillance.China has a vast territory,numerous mountains and rivers.Bridge is the national infrastructure that has become an indispensable part of people's production and life for thousands years.The safety of the bridge has always been dependent on manual testing and bridges that likely to get safety problems are usually built on mountains.There is a lot of risks in artificial detection of bridge,and it is easy to appear one-sided,leaving a lot of security risks.The Application of unmanned aerial vehicle bridge detection can not only easily avoid the risk of manual detection,but also a more comprehensive inspection of the bridge.In this paper,the bridge crack detection of UAV is studied in the laboratory environment.UAV bridge crack detection involves many key parts,such as UAV flight,UAV picture processing,bridge model establishment,crack extraction and evaluation,In the crack image processing section,this paper uses the image mosaic technique to obtain the panoramic image of the bridge.The feedback mechanism is added between the rough matches and fine matches to minimize the cumulative error in the mosaic of the image.The results show that the feedback mechanism can control the error well to obtain high quality panoramic images.UAV flight and bridge models building are based on RGBDSLAM for research and experimentation.Firstly,this paper use the depth image and color image to generate the real point cloud model;Secondly,according to the relationship between the point cloud,this paper can build a visual odometer,posture map optimization method is applied to stabilize the entire map of the building.Finally,the loopback detection mechanism based on the word pocket model is also a good way to reduce the errors in the map construction,greatly enhancing the credibility of the map.As the dense map in depth of the camera reconstruction,UAV establish a three-dimensional model of the bridge during autonomous flight,as well as the flight path.In bridge crack extraction part,Otsu technology is used to obtain the full information of bridge crack.The maximum variance projection technique leads to the main direction of the crack,and gives the cracking of the crack as a whole.Gini coefficient analyzes the trend of crack cracking and provides a data basis for dynamic analysis of fractures.The degree of polymerization mainly calculates the width of the crack information.The experimental results show that these three measurement coefficients can effectively distinguish the cracks and give the key information of the cracks,and leave some characteristics information for the later bridge classification.