The accuracy of low probability of intercept (LPI) radar waveform recognition is an important and challenging problem in electronic warfare. Aiming at the problem of the difficulty in feature extraction and the low recognition rates of the LPI radar signal under a low signal-to-noise ratio, and inspired by the symmetry theory, we propose a new approach for the LPI radar signal recognition method based on a dual-channel convolutional neural network (CNN) and feature fusion. Our new approach contains three main modules: the preprocessing module that converts the LPI radar waveforms into two-dimensional time-frequency images using the Choi–Williams distribution (CWD) transformation and performs image binarization, the feature extraction module that extracts different features obtained from the images, and the recognition module that utilizes a multi-layer perceptron (MLP) network to fuse these features and distinguish the type of LPI radar signals. In the feature extraction module, a two-channel CNN model is proposed that extracts Histogram of Oriented Gradients (HOG) features and deep features from time-frequency images, respectively. Finally, the recognition module recognizes the radar signals using a Softmax classifier based on the fused features from two channels. The experimental results from 12 types of LPI radar signals prove the superiority and robustness of the proposed model. Its overall recognition rate reaches 97% when the signal-to-noise ratio is −6 dB.

Time‐frequency analysis based on Wigner‐Ville distribution (WVD) plays a significant role in analysing non‐stationary signals, but it is susceptible to interference from crossterms (CTs) for multi‐component signals. To address this issue, a novel WVD enhancement method based on generative adversarial networks (namely WVD‐GAN) is proposed, to achieve highly‐concentrated time‐frequency (TF) representation. Specifically, a deep feature extraction module is designed with multiple residual connections in the generator of WVD‐GAN to leverage the latent information encoded in the shallow representations. Meanwhile, a simple and effective attention module is introduced to enhance auto‐term features. Moreover, a multi‐scale discriminator is proposed based on dilated convolutions to guide the generator to reconstruct high‐resolution TF images by discriminating CT. Finally, a comparative analysis is provided to demonstrate the effectiveness and robustness of the proposed method on different simulated and real‐life datasets. Extensive experiments demonstrate that the proposed method outperforms several state‐of‐the‐art methods.

Radar signal recognition plays a crucial role in modern electronic reconnaissance systems. With the increasing complexity of electromagnetic environments, radar signals are susceptible to noise interference under low signal-to-noise ratio (SNR) conditions, posing a challenge to accurate radar signal recognition. To address this issue, we propose a multilayer decomposition denoising empowered convolutional neural network (CNN) for radar signal recognition. Specifically, the original radar signals are first processed by multilayer decomposition denoising, which consists of variational mode decomposition (VMD), local mean decomposition (LMD), and wavelet thresholding (WT) in sequence, termed as VMD-LMD-WT. Then we use the Choi-Williams distribution (CWD) to convert the denoised signals into time-frequency images (TFIs). Finally, an improved CNN with dilated convolution is employed for radar signal recognition. Experiments demonstrate that the multilayer decomposition denoising method effectively improves the SNR of the original signal, which is beneficial for the subsequent recognition task. The recognition accuracy is improved by about 11% compared to that directly using the original signal. Furthermore, compared to the current network models, the proposed CNN network can efficiently extract the signal features and improve the detection accuracy of low probability of intercept (LPI) radar signals. The recognition accuracy reaches 75.3% for 12 signals when the SNR is low to -14dB.

Anomalous road manhole covers pose a potential risk to road safety in cities. In the development of smart cities, computer vision techniques use deep learning to automatically detect anomalous manhole covers to avoid these risks. One important problem is that a large amount of data are required to train a road anomaly manhole cover detection model. The number of anomalous manhole covers is usually small, which makes it a challenge to create training datasets quickly. To expand the dataset and improve the generalization of the model, researchers usually copy and paste samples from the original data to other data in order to achieve data augmentation. In this paper, we propose a new data augmentation method, which uses data that do not exist in the original dataset as samples to automatically select the pasting position of manhole cover samples and predict the transformation parameters via visual prior experience and perspective transformations, making it more accurately capture the actual shape of manhole covers on a road. Without using other data enhancement processes, our method raises the mean average precision (mAP) by at least 6.8 compared with the baseline model.