中圖分類號： TP391.41
文獻標識碼： A
DOI： 10.19358/j.issn.2096-5133.2021.02.005
引用格式： 柳勝超，王夏黎，張琪，等. 數(shù)字圖像處理在橋梁結(jié)構(gòu)變形檢測的應(yīng)用研究[J].信息技術(shù)與網(wǎng)絡(luò)安全，2021，40(2)：24-32.

Application research of digital image processing in deformation detection of bridge structures

Abstract： In view of the structural deformation of large bridges during construction and operation, there is currently a lack of automated, high-frequency, real-time, long-term and accurate detection methods. Based on the theory of digital image processing and deep learning, this paper proposes a non-contact detection method suitable for large-scale bridge structure deformation, and uses this method to develop a system that can simultaneously dynamically monitor multiple target structures of the bridge. This method firstly obtains dynamic video sequence images of the bridge structure through high-resolution photography equipment; secondly, it preprocesses the image to remove the influence of external factors such as weather on the image; then it extracts the image ROI to determine the specific bridge structure to be processed; finally, the YOLOv3 algorithm is improved and combined with the improved SURF algorithm to realize the deformation detection of the bridge structure. Experimental results show that the detection speed of the algorithm is between 20 fps and 30 fps, when the target distance is 100 m, the detection accuracy of the algorithm is within 0.3 mm, and the detection accuracy is high, which effectively reflects the deformation of the bridge structure.

Key words : software engineering；bridge structure deformation；digital image processing；remote detection；deep learning；SURF algorithm