The reconstruction of the self-luminous image of a flame through deep learning can inform research on the characteristics of combustion of a scramjet. In this study, the authors propose a residual network model based on the channel and spatial attention mechanisms to reconstruct the self-luminous image of a flame from schlieren images of the flow field of a scramjet. We compare the reconstruction-related performance of single-path and dual-path models under different conditions. The channel and spatial attention mechanisms enable the model to focus on important feature-related information, and the residual connection prevents gradient disappearance to improve the capability of the model for generalization. The proposed method was tested through a supersonic combustion experiment in a ground wind tunnel under different equivalence ratios, and data on the flow field of the combustion chamber and the evolution of the flame were recorded as a dataset. A number of experiments as well as subjective and objective analyses were subsequently carried out on this dataset. The results show that the effect of reconstruction is consistent with the original image of the flame, and the geometric characteristics of the flame are accurately reconstructed.

Extraction of the wave structure of the flow field in the combustor of the scramjet is important for main flow control and performance evaluation of the scramjet. In this study, a deep learning-based method based on the fully convolutional network with 8-pixel stride is proposed to segment the schlieren image to extract the wave structure. First, use a residual neural network with 34 layers as the backbone network to extract features, which ensures highly efficient learning through residual blocks to extract multi-dimension semantic information. Second, dilated convolution is utilized to expand the receptive fields of deepened layers to obtain high-dimensional features and increase the degree of aggregation of contextual information contained in the high-dimensional features. Finally, the channel and spatial attention module are introduced to the decoding stage to enable the model to focus on key information to improve the segmentation accuracy. A large number of experiments are carried out on a dataset of schlieren images of the flow field in the combustor of scramjet that were compiled by the authors. The proposed method recorded higher values of the pixel accuracy, recall, intersection over union, and F1 score than compared methods, with values of 78.47%, 83.81%, 67.51%, and 80.32%, respectively. This method can effectively complete the wave structure extraction and provide important basic support for related research work of scramjet.

The rapid and accurate reconstruction of the supersonic combustor flow field is of great significance for sensing and predicting the combustion state. Existing deep learning methods pay less attention to the convergence speed of flow field reconstruction, which results in longer training and prediction times for the models. This study proposes a method for reconstructing the flow field in supersonic combustor by combining a reduced-order model based on proper orthogonal decomposition (POD) with a multi-branch convolutional neural network. This method first analyzes the effectiveness of POD reconstruction. Then, based on the wall pressure data of the supersonic engine combustor, it performs flow field image reconstruction. Finally, through error calculation and gradient updating with low-resolution principal component flow field shadow images obtained from the POD algorithm, the high-precision and efficient prediction of flow field images is achieved. Different equivalence ratio hydrogen fuel combustion experiments were conducted in a pulsed combustion wind tunnel with an incoming flow Mach number of 2.5. The learning model was trained and tested using the dataset obtained from these experiments. Numerous experiments demonstrated that the model can effectively reconstruct the wave structures of complex flow fields. Multiple evaluation indicators indicated that the reconstructed flow field of the combustor shows good agreement with that obtained from ground wind tunnel testing. Furthermore, after introducing the POD dimensionality reduction model, the training time was reduced by 32.03%, effectively improving the training time complexity of the model.

Timely detection of defects is essential for ensuring safe and stable operation of concrete buildings. Automatic segmentation of concrete buildings’ surfaces is challenging due to the high diversity of crack appearance, the detailed information, and the unbalanced proportion of crack pixels and background pixels. In this work, the Double Feature Pyramid Network is designed for high-precision crack segmentation. Our work reached the state-of-the-art level in crack segmentation, with key contributions outlined as follows: firstly, considering the diversity of crack shapes, the network constructs a feature pyramid containing three feature extraction backbones to extract the global feature map with three scale input images. In particular, due to the biggest challenge being too much single-pixel crack area, the targeted feature pyramid based on the high-resolution is added to extract adequate shallow semantic information. Lastly, designing a cascade feature fusion unit to aggregate the extracted multi-dimensional feature maps and obtain the final prediction. Compared with existing crack detection methods, the superior performance of this method has been verified based on extensive experiments, with Pixel Accuracy of 65.99%, Intersection over Union of 44.71%, and Recall of 62.95%, providing a reliable and efficient solution for the health monitoring and maintenance of concrete structures. This work contributes to the advancement of research and practical applications in related fields, offering robust support for the monitoring and maintenance of concrete structures.

The essence of Chinese calligraphy inheritance resides in calligraphy education. However, it encounters challenges such as a scarcity of calligraphy instructors, time-consuming and inefficient manual assessment methods, and inconsistent evaluation criteria. In response to these challenges, this paper introduces a deep learning-based automatic calligraphy evaluation model. Initially, hard-pen handwriting samples from 100 volunteers were collected and preprocessed to create a dataset consisting of 4800 samples, along with the corresponding label files for hard-pen calligraphy evaluation. Subsequently, YOLOv5 was utilized for region detection and character recognition on the evaluation samples to obtain the corresponding standard samples. Lastly, a Siamese metric model, with VGG16 as the primary feature extraction submodule, was developed for hard-pen calligraphy evaluation. To improve feature extraction and propagation, a transformer structure was introduced to extract global information from both the evaluated and standard samples, thereby optimizing the evaluation results. Experimental results demonstrate that the proposed model achieves a precision of 0.75, recall of 0.833, and mAP of 0.990 on the hard-pen calligraphy evaluation dataset, effectively realizing automatic calligraphy evaluation. This model presents a novel approach for intelligently assessing hard-pen calligraphy.