High-precision bridge apparent disease identification method based on computer vision

• Main Authors: NI YOUHAO, WAN YAHUA, MAO JIANXIAO, XIE YISHUN, WEN XUEHUA, PANG ZHENHAO, WANG FEIQIU, WANG HAO, ZHAO KAIYONG
• Format: Patent
• Language: Chinese; English
• Published: 24.11.2020
• Subjects: CALCULATING; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention discloses a high-precision bridge apparent disease identification method based on computer vision. The method comprises an image preprocessing stage, a simulation image generation stageand an apparent disease identification stage. In the image preprocessing stage, Gaussian filtering and image pixel equalization processing are carried out on a training set containing disease images.In a simulation image generation stage, a data distribution mode of a generative adversarial network learning training set is adopted to generate a simulation image, so that the data scale of the training set is increased. And in the disease identification stage, the increased disease training set is adopted to train the YOLO model, and the trained model is used for bridge apparent disease identification. When the scale of the training data set is small, the scale of the neural network training data set is expanded by adopting the generative adversarial network, and the precision of apparent disease identification of t