Remote sensing scene classification remains challenging due to the complexity and variety of scenes. With the development of attention-based methods, Convolutional Neural Networks (CNNs) have achieved competitive performance in remote sensing scene classification tasks. As an important method of the attention-based model, the Transformer has achieved great success in the field of natural language processing. Recently, the Transformer has been used for computer vision tasks. However, most existing methods divide the original image into multiple patches and encode the patches as the input of the Transformer, which limits the model’s ability to learn the overall features of the image. In this paper, we propose a new remote sensing scene classification method, Remote Sensing Transformer (TRS), a powerful “pure CNNs → Convolution + Transformer → pure Transformers” structure. First, we integrate self-attention into ResNet in a novel way, using our proposed Multi-Head Self-Attention layer instead of 3 × 3 spatial revolutions in the bottleneck. Then we connect multiple pure Transformer encoders to further improve the representation learning performance completely depending on attention. Finally, we use a linear classifier for classification. We train our model on four public remote sensing scene datasets: UC-Merced, AID, NWPU-RESISC45, and OPTIMAL-31. The experimental results show that TRS exceeds the state-of-the-art methods and achieves higher accuracy.

Road crack detection is an important task for road safety and road maintenance. In the past, people made use of manual detection methods and tried to use computer vision to detect crack. The most prominent feature in recent years is the use of deep learning. However, there is no good deep learning method for road crack detection under noise. This challenge is faced bravely. First, a noise crack dataset is proposed, consisting of multiple noise crack images which is called NCD. Then, an adaptive bilateral filtering algorithm is developed, which can reduce the influence of noise. Finally, a new crack detection network with two new modules is designed. In the end, it is found that all the parts have promoting effects on crack detection under noise. Compared with other state‐of‐the‐art methods, this method performs better, especially in road crack detection under noise. When evaluating the well‐known crack500 test set, ODS F‐measure of 0.628 is achieved. Besides, this method is also evaluated in another five datasets. Significantly, ODS F‐measure of 0.545 is achieved, 4.0% higher than state‐of‐the‐art on GAPs384.