Bing Li

Military Camouflage Target Detection (MCTD) is a special object detection task that aims to detect military camouflaged targets in the wild. In the challenging MCTD task, the confusing appearance and contours of military camouflaged targets often lead to the poor performance of existing methods. In this study, we propose an end-to-end Military Detection Transformer (MilDetr) for MCTD. We introduce two improvements to enhance the model’s performance. First, we employ the Reverse Features Feed Forward Neural Network (R3FN) for local information aggregation in the encoder of MilDetr. In addition, the Fusion Previous Query (FPQ) module is utilized for multi-stage query feature fusion in the decoder of MilDetr. To overcome data limitations for MCTD, we build two simulation military camouflaged target datasets called MilDet and MilCls. The ablation experiments on MilDet reveal the effectiveness of our improvements. Experimental results demonstrate that MilDetr obtains 95.6 AP on MilDet. Furthermore, MilDetr obtains 96.4 AP on MilDet with the pre-trained weights on ImageNet and MilCls. Compared with other object detectors, MilDetr achieves end-to-end military camouflaged target detection with superior performance.

In this study, an improved end-to-end framework for instance segmentation of military camouflaged targets, referred to as MilInst, is proposed. The framework builds upon SparseInst method developed by Cheng et al . [27]. Several improvements are introduced to enhance the model’s performance. First, Receptive Field Enhancement Module (RFEM) is employed to capture broader contextual information. Additionally, Feature Merging Module (FMM) is utilized to eliminate feature noise through the implementation of a matrix decomposition method. Furthermore, a novel linear dynamic bipartite matching approach is proposed, facilitating a smooth transition from one-to-one matching to one-to-many, and eventually achieving accurate to one-to-one matching. Experimental results demonstrate the effectiveness of MilInst algorithm. Comparisons with a selected real-time instance segmentation baseline model reveal superior performance, with MilInst framework achieving the highest mean Average Precision (mAP) index of 84.8% on a self-made dataset.

In the current battlefield space, with the massive application of electromagnetic equipment, the electromagnetic environment in the battlefield space tends to be complex, which can lead to the electromagnetic equipment and personnel in the battlefield space receiving interference from the electromagnetic environment signal. To protect the safety of personnel and equipment quality, it is necessary to evaluate the complexity of the electromagnetic environment signal research, to use the corresponding measures. However, there is still little research related to the evaluation of the complexity of electromagnetic environmental signals. In this paper, a feature extraction method for electromagnetic environmental signals based on adaptive multiscale morphological gradient filtering and a nonnegative matrix factorization algorithm is proposed. First, the electromagnetic environment signal is filtered by AMMG, and then the filtered signal is processed by NMF for feature extraction. Finally, the complex electromagnetic environment signals after feature extraction are evaluated and classified by the SVM method. The results show that the evaluation results have good classification accuracy, and this paper provides an effective feature extraction method for the complexity of electromagnetic environment signals.

In this paper, a feature extraction method for evaluating the complexity of the Electromagnetic Environment (EME) of the photovoltaic power station is presented by using logarithmic morphological gradient spectrum (LMGS) based on the mathematical morphological theory. We use LMGS to evaluate electromagnetic environment signals. We also explored the impact of structure element (SE) on the MS, MGS, and LMGS. Three types of SE, mean the line SE, square SE and diamond SE, are utilized and compared for computing the LMGS. EME signals with four complexity degrees are simulated to evaluate the effectiveness of the presented method. The experimental results have shown that the feature extraction scheme proposed in this paper is a reasonable method to classify the complexity of EME.

Bearing fault signal analysis is an important means of bearing fault diagnosis. To effectively eliminate noise in a fault signal, an adaptive multiscale combined morphological filter is proposed based on the theory of mathematical morphology. Both simulation and experimental results show that the adaptive multiscale combined morphological filter can remove noise more thoroughly and retain details of the fault signal better than the dual-tree complex wavelet filter, traditional morphological filter, adaptive singular value decomposition method (ASVD), and improved switching Kalman filter (ISKF). The adaptive multiscale combined morphological filter considers both positive and negative impulses in the signal; therefore, it has strong adaptability to complex noise in the environment, making it an effective new method for bearing fault diagnosis.

In order to solve the problem that the accuracy is not high when the traditional fuzzy morphological associative memories network (FMAM) is used to classify samples, based on quantum superposition and collapse, the quantum fuzzy morphological associative memories network (QFMAM) is proposed. In QFMAM, the structural element is constructed by the qubit system and the qubit probability represents corresponding membership degree to obtain adaptive structural elements. The samples are preprocessed first and then classified to ensure high classification accuracy. QFMAM, FMAM, support vector machine (SVM) and naive Bayes classifier (NBC) are used to classify simulation data and experimental signals of rolling bearings respectively. Comparing the performance of the four classifiers, it is obvious that the training time of QFMAM is far less than SVM and NBC, and the classification accuracy of the test samples by QFMAM is higher than that by the other three classifiers. QFMAM is very suitable for intelligent diagnosis of rolling bearing fault.

Wavelet transform is one of the most desirable tools for depressing noise. However, the traditional linear wavelets are not always suitable for any real world signals with strong background noises. In this work, we present a new morphological wavelet, named averaged dilation‐erosion morphological wavelet (ADEMW), for depressing the noise in signals of firing shock force on the shoulder. Simulated signals with different SNRs are generated to evaluate and compare the proposed new wavelet scheme with the traditional linear wavelet and another two morphological wavelets presented in literature. Experimental results reveal that the presented ADEMW gives the most promising noise suppression performance. Then, the ADEMW is employed to process the real‐world signals acquired from a firing shock force testing system. Processing results demonstrate that the ADEMW also outperforms another three wavelets obviously for depressing the strong background noise in the signals of firing shock force on the shoulder. The main impulsive components in the firing shock force can be clearly detected for analyzing the impacts on shoulder during the shooting process. The presented ADEMW scheme has provided a novel desirable tool for analyzing the complicated signals with strong noise.

Feature extraction and faults classification are the two most significant issues involved in the field of mechanical fault diagnosis problems. In this work, we address these two problems using mathematical morphology and non-negative matrix factorization. In particular, we present a novel engine fault diagnosis scheme utilizing the averaged multi-scale morphological filter to enhance the vibration signals, non-negative matrix factorization to characterize the signals, and a constructive morphological neural network to classify the engine operating states. Eight engine running states including the healthy state and seven defective states are tested in an engine experiment rig to evaluate the presented fault diagnosis scheme. Conventional feature extraction methods as well as classifiers popularly used in the literature are also employed as a comparison. The experimental results indicate the proposed approach to be an effective and efficient scheme for detection of the intelligent faults of engines.