The impact of crack diseases on roads and other infrastructure will gradually increase with the passage of time,so it is necessary to detect pavement cracks.There are problems in the similar gray values and low detection accuracy of asphalt pavement crack diseases.So the study proposes a method of asphalt pavement crack identification based on convolutional neural network VGG16;constructs the asphalt pavement crack data set;divides them into training set,verification set and test set;conducts pre-processing and data enhancement through geometric transformation such as cutting and rotation;trains with VGG16 model;and predicts whether there are cracks in asphalt pavement.The results show that under the condition of the small number of asphalt pavement crack data sets and the complex environment,the detection accuracy is better,which has certain engineering application value for asphalt pavement crack detection.
维普
万方数据