UAV detection can overcome the limitation of space, greatly improve the efficiency of bridge detection, save detection cost and ensure the safety of personnel.
At present, the research on UAV bridge detection in China is still in a separate development state, that is “man in the loop based on UAV semi-automatic bridge detection”. UAV itself does not have the function of independent identification structure and automatic route planning, and only serves as a data acquisition medium. The research focuses on the image post-processing technology (such as object classification, detection, segmentation, 3D reconstruction, etc.), and its essence is image processing based on computer vision and deep learning, which converts the captured pictures and videos into indicators that bridge engineers can understand or data that can support decision-making.
Automatic UAV bridge detection requires independent route planning, automatic data collection and automatic data analysis, and the analysis results directly serve the decision-making. This is the future trend.
Nowadays, North Carolina is in the forefront of bridge detection technology practice in the United States, and has become the focus of bridge detection industry. They have successively completed the laser mapping operation of the bridge and the UAV over the horizon imaging operation.
North Carolina proving ground
These are two meaningful practices. First of all, the laser mapping of Bassnett bridge completed at the end of September this year is a thorough scanning of the whole picture of the bridge (Marc basnightbridge) with trillions of laser beams to observe the strain and structural damage of the bridge concrete. Then in early October, the Federal Aviation Administration (FAA) approved the North Carolina Department of transportation (ncdot) to use UAVs for bridge inspection beyond visual range, which officially opened a new era of structural safety inspection by UAVs.
It is not uncommon to use UAV to detect bridge structural damage. Since 2013, the United States has been trying to use UAV detection in the state transportation department. However, due to signal and safety reasons, UAV has not been ideal in the processing of beyond visual range and dead angle of bridge structure. Only when the coverage of UAV is extended beyond the operator’s line of sight, can more high-resolution images be collected in places that are difficult for human beings to reach, and the analysis of bridge structure can be complete and standard, which is conducive to the later work of input, analysis and comparison.
Eric Boyette, head of North Carolina’s Department of transportation, is very pleased that these two detection technologies can be first applied in North Carolina. “Vehicle borne laser detection and over the horizon UAV automatic detection are two important technical practices, and the advantage of both vehicles lies in their minimal impact on traffic,” he said. Especially for UAV detection, the current task is to ensure the integrity and safety of UAV detection process, and the recognition ability of other details will be improved in the future. This is a brand new system that will enable us to complete the remote inspection of the bridge. “
Some of the applications for the project were developed by skydio, a maker of drones. It is characterized by airborne artificial intelligence operating system and power system, which can make the aircraft avoid obstacles (such as bridge substructure and lower truss) in the absence of reliable GPS signal, so as to complete the whole process of detection task safely and completely. This is also one of the difficulties of automatic UAV detection.
A breakthrough in technology and cost
UAV detection has some traditional difficulties. The first is the signal of UAV. In practical operation, UAV usually navigates through GPS, but GPS signal is easily interfered by external conditions, which leads to no signal and unstable signal, resulting in unable to complete the detection. The second is the automatic operation of man-machine. Commercial UAV flight will be affected by weight, wind field and other factors, even professional pilots will be difficult to control when driving. The UAV used for bridge detection needs to have certain particularity, which requires customized aircraft loading AI autopilot technology. The third is the acquisition of laser imaging and UAV image data and three-dimensional imaging technology, which has special requirements for flight pattern and focusing accuracy. The fourth is about the identification of different types of bridge damage, mainly for the detection of concrete cracks and steel structure damage. Whether in the air or on the road, the laser scanning imaging should be able to directly connect with the preset bridge health analysis system, and the analysis process can be semi manual or automatically completed by the computer.
The UAV in this test uses the most advanced AI automatic flight engine in the world. This technology can make UAV complete automatic and safe navigation in 360 ° full obstacle environment, even in the area without GPS signal. In terms of imaging transmission, a 150 meter steel truss bridge can complete all imaging transmission operations in two hours and return them to the database for synchronous analysis.
In terms of cost, over the horizon automatic UAV detection technology will replace manual observers or UAV drivers, which will further reduce the cost of bridge detection. According to the calculation of AASHTO, compared with traditional detection methods, the cost of UAV detection can be reduced by 75%, which is mainly reflected in the significant reduction of labor cost and time cost. Besides the saved social interference cost (such as bridge closure and circuit breaking), the comprehensive cost of UAV detection will be lower.
On the market side, the head of the North Carolina Department of transportation said it was an “achievement that can change the rules of the game.”. Once this technology is further mature, its market will be 20000 bridges in the state and 600000 bridges in the United States. It is hopeful that dangerous slings and expensive labor costs will be completely abandoned, and it will pave the way for future bridge and infrastructure inspection.
Editor in charge: GT