Bridge bottom surface crack detection method based on unmanned aerial vehicle vision

Abstract

The invention discloses a bridge bottom surface crack detection method based on unmanned aerial vehicle vision. The method comprises the following steps: shooting to-be-detected bridge span blocks toobtain a plurality of RGB (Red-Green-Blue) color images by an unmanned aerial vehicle carrying a high-definition zoom camera according to a preset flight path, a single shooting area and a shooting sequence, consecutively numbering the RGB color images, and preprocessing the color images into gray level images to be saved; performing image enhancement and noise reduction filtering on the gray level images; processing the images into binary images by utilizing MATLAB (Matrix Laboratory) software, and segmenting clear crack images; performing projection computation in a multi-axis direction by using a projection method so as to recognize corresponding crack types, and establishing a three-layer BP neural network; and evaluating the crack disease development conditions after all the cracks are recognized. The bridge bottom surface crack detection method disclosed by the invention has the crack classification and recognition accuracy of 90% or higher and also has the advantages of degree of automation, high efficiency and the like, and the feasibility and effectiveness in the process of detecting the bridge bottom surface cracks are verified.

Description

technical field [0001] The invention relates to the field of bridge bottom crack detection, in particular to a bridge bottom crack detection method based on drone vision. Background technique [0002] The life cycle of a bridge is usually divided into three stages: 1. Design stage; 2. Construction stage; 3. Operation, management and maintenance stage. In the life cycle of a bridge, the construction phase usually only accounts for 10%, while the maintenance management phase accounts for 90% of the bridge life cycle. The hidden dangers of bridge safety cannot be ignored, and bridge management, maintenance and testing are related to the safety of people's lives and property. [0003] The crack and disease detection of bridges is mainly based on manual on-site investigation. The following four methods are commonly used: [0004] 1) Inspectors use telescopes and other equipment to conduct long-distance observation of bridges; [0005] 2) The inspectors use the bridge inspectio...

AI Technical Summary

Problems solved by technology

[0008] These traditional detection methods have corresponding defects, such as large workload of long-distance observation, slow speed, low efficiency, easy to cause eye fatigue of inspectors, high missed detection rate, unacceptable ground conditions under the bridge, etc.; bridge inspection vehicle cost Expensive, not easy to popularize, occupying bridge deck lanes, posing a threat to driving s

Images