Developing a new deep learning CNN model to detect and classify highway cracks

Purpose: Testing the capabilities/accuracies of four deep learning pre-trained CNN models to detect and classify types of highway cracks, as well as, developing a new CNN model to maximise the accuracy at different learning rates.
Design/methodology/approach: a sample of 4,663 images of highway cracks were collected and classified to three categorises of cracks, namely, vertical cracks’ ‘horizontal and vertical cracks’ and ‘diagonal cracks’, subsequently, using ‘Matlab’ to classify the sample to training (70%) and testing (30%) to apply the four deep learning CNN models and compute their accuracies. After that, developing a new deep learning CNN model to maximise the accuracy of detecting and classifying highways cracks and testing the accuracy using three optimisation algorithms at different learning rates.
Findings: the accuracies result of the four deep learning pre-trained models are above the averages between top-1 and top-5 and the accuracy of classifying and detecting the samples exceeded the top-5 accuracy for the pre-trained AlexNet model around 3% and by 0.2% for the GoogleNet model. The accurate model here is the GoogleNet model as the accuracy is 89.08% and it is higher than AlexNet by 1.26%. While the computed accuracy for the new created deep learning CNN model exceeded all pre-trained models by achieving 97.62% at learning rate 0.001 using Adam’s optimisation algorithm.
Practical Implications: The created a deep learning CNN model will enable users (e.g., highways agencies) to scan a long highway and detect types of cracks accurately in a very short time compared to traditional approaches.
Originality/value: A new deep learning CNN-based highway cracks detection was developed based on testing four pre-trained CNN models and analyse the capabilities of each model to maximise the accuracy of crack detection based on the proposed CNN.