Analysis of bridge foundation pile detection based on convolutional neural network model

Abstract In order to quickly and accurately detect the instability of foundation piles on concrete Bridges, an improved convolutional neural network based image recognition method for bridge pile foundations was proposed. improve quality, entropy threshold is used process image, two-channel designed extract features fully. The traditional Relu activation function avoids model underfitting. Support vector machine (SVM) replace Softmax classifier computing efficiency. experiment shows that has significantly rate real foundation. No matter how many degrees rotated, in this article always maintains a high rate, does not fluctuate much, indicating algorithm good robustness rotation translation. summary, average 5 groups 96.26%. feasibility identifying proved.