The low altitude, slow speed and small size object which we call LSS-object for short, such as small UAV (unmanned aerial vehicles) have become a hot issue of air defense security, which is difficult to detect and ide...
详细信息
The low altitude, slow speed and small size object which we call LSS-object for short, such as small UAV (unmanned aerial vehicles) have become a hot issue of air defense security, which is difficult to detect and identify accurately from the image. In this paper, aiming at the problem of LSS-object detection under noise environment, the detection method based on deep learning is proposed. Firstly, a standard training dataset consisting 5 classes of typical objects is constructed. Then, the standard dataset is augmented with noise of different intensity. Finally, YOLO v3 algorithm is used to form a LSS-object detection system which can adapt to environment noise. The training and detection experiments were carried out on the GPU server. After only using the noise-free dataset for training, the mAP(mean Average Precision) of the noise-free test set detection reached 81.07%, but the mAP decreased to 20.68% when the noise variance was 0.03. After adopting the mixed training strategy of the dataset with noise variance of 0.01 and noise-free data, the mAP for the test set detection with noise variance of 0.03 was increased to 70.61%, and the mAP still reached 79.85% in noise-free test set detection. The experiment results show that the mixed training strategy can greatly improve the accuracy in the noisy images detection while maintaining a higher accuracy in noise-free images.
Constant false alarm rate (CFAR) is a traditional method for radar target detection. A lot of relevant researches assume that the sea clutter obeys the K distribution. However, with the improvement of the radar resolu...
详细信息
Constant false alarm rate (CFAR) is a traditional method for radar target detection. A lot of relevant researches assume that the sea clutter obeys the K distribution. However, with the improvement of the radar resolution and the increase of the grazing angle, the distribution of sea clutter gradually deviates from the K distribution. Although good fitting results can be obtained through the more complicated sea clutter models such as KA distribution, K+Rayleigh distribution and so on, the computational efficiency of the detection algorithm is reduced obviously. For computational complexities are mainly caused by non-closed expression, A hybrid sea-clutter distribution (KK distribution) with closed-form expression for high-resolution and high grazing angle situation is proposed, which fits the measured data. Inspired by this, a new hybrid sea-clutter distribution (KR distribution) with 4 parameters is proposed. A new method of parameter estimation for KR distribution is given, that is, passing-parameter estimation method based on hybrid distribution. The results show that the fitting effect of KR distribution is better than that of both K distribution and KK distribution in case of high resolution and large grazing angle. And the CA-CFAR based on KR distribution of passing-parameter estimation performs relatively better in the aspect of target detection.
This paper considers the problem of tracking a moving target using the time difference of arrival (TDOA) and frequency difference of arrival (FDOA) measurements obtained at two unmanned aerial vehicles (UAVs) with var...
详细信息
This paper considers the problem of tracking a moving target using the time difference of arrival (TDOA) and frequency difference of arrival (FDOA) measurements obtained at two unmanned aerial vehicles (UAVs) with varying baseline. Accumulation time is necessary because the emitter position cannot be estimated at each emission when there are only two UAVs as sensors. And the target position and velocity estimation accuracy often suffer from low convergence in conventional formation flight mode no matter what nonlinear filtering algorithm is used. Based on the analysis of the influence of different flight modes on the positioning performance, a moving target tracking system from two UAVs with varying baseline is proposed. The performance of unscented Kalman filter (UKF) under proposed system is analysed and compared with the Cramer-Rao lower bound (CRLB). Simulation results show that the proposed system can speed up the convergence.
Stereo matching is one of the key steps of 3D reconstruction based on binocular vision. In order to improve the convergence speed and accuracy in 3D reconstruction based on binocular vision, this paper adopts the comb...
详细信息
Stereo matching is one of the key steps of 3D reconstruction based on binocular vision. In order to improve the convergence speed and accuracy in 3D reconstruction based on binocular vision, this paper adopts the combination method of polar constraint and ant colony algorithm. By using the line constraint to reduce the search range, an ant colony algorithm is used to optimize the stereo matching feature search function in the proposed search range. Through the establishment of the stereo matching optimization process analysis model of ant colony algorithm, the global optimization solution of stereo matching in 3D reconstruction based on binocular vision system is realized. The simulation results show that by the combining the advantage of polar constraint and ant colony algorithm, the stereo matching range of 3D reconstruction based on binocular vision is simplified, and the convergence speed and accuracy of this stereo matching process are improved.
In map generalization, a technical difficulty and key to expressing water elements is river classification, which has high requirements for accuracy and calculation speed. In the study of river classification, the cur...
详细信息
In map generalization, a technical difficulty and key to expressing water elements is river classification, which has high requirements for accuracy and calculation speed. In the study of river classification, the current hierarchical algorithm is not quite efficient and lacks classification of ring rivers. Based on this, this paper establishes the attributes of river classification, its geometric rules and ring rules, and classifies different ring types for the existing in river system. When calculating river's longest path, we use the heuristic algorithm for iterative calculation to automatically conduct river classification. The experimental results show that this algorithm can achieve good results for the ring river system and improve the calculation efficiency and accuracy of river system classification.
This paper presents a novel three-component scattering power decomposition of polarimetric SAR data. There are two problems in three-component decomposition method: volume scattering component overestimation in urban ...
详细信息
This paper presents a novel three-component scattering power decomposition of polarimetric SAR data. There are two problems in three-component decomposition method: volume scattering component overestimation in urban areas and artificially set parameter to be a fixed value. Though volume scattering component overestimation can be partly solved by deorientation process, volume scattering still dominants some oriented urban areas. The speckle-like decomposition results introduced by artificially setting value are not conducive to further image interpretation. This paper integrates the results of eigen-decomposition to solve the aforementioned problems. Two principal eigenvectors are used to substitute the surface scattering model and the double bounce scattering model. The decomposed scattering powers are obtained using a constrained linear least-squares method. The proposed method has been verified using an ESAR PolSAR image, and the results show that the proposed method has better performance in urban area.
Weiss39;s Direct position determination (WDPD) method compensate the received signals of the M stations to the transmitted signals by the Delay-Doppler operator, and Delay-Doppler operator is computed M times to com...
详细信息
Weiss's Direct position determination (WDPD) method compensate the received signals of the M stations to the transmitted signals by the Delay-Doppler operator, and Delay-Doppler operator is computed M times to compensate the time delay and Doppler shift for every search point, and the dimension of the Delay-Doppler operator is related to the size of data points N. And Delay-Doppler operator is equal to four matrices multiplying continuously, whose size is NxN, and matrix multiplication needs lots of time and memory, limiting the size of N. This paper deduces the DPD that don't need to compensate to the transmitted signals. In practice, any one of M stations maybe is the reference station, and the signal of the reference station can compensate to the receiving signals of other stations by Delay-Doppler operator, and the objective function of the DPD is given. Through the DPD, Delay Doppler operator only is calculated M-1 times for every search point, and reduce the computation complexity, especially M is small. And for DPD algorithm, there are a lot of redundant computations in the Delay Doppler operator, so the fast algorithm of DPD (FDPD) is proposed. The experimental results verify that DPD reduce the computation complexity and FDPD saves the computation time and maintain the performance of the DPD algorithm simultaneously.
Filter plays a vital role in electronics and communication system. Filters are used to pass the desired frequency signals and rejects undesired frequency signal. There are several applications where digital filtering ...
详细信息
The proceedings contain 70 papers. The topics discussed include: a novel simulation system oriented description language of worm propagation;a novel micro-motion target parameter estimation algorithm in the situation ...
The proceedings contain 70 papers. The topics discussed include: a novel simulation system oriented description language of worm propagation;a novel micro-motion target parameter estimation algorithm in the situation of low SNR;space-time-polarization adaptive antenna arrays for GNSS receivers;an improved training sample set construction method;time-singularity power spectral distribution and its application in the analysis of radar sea clutter;extraction of disturbance signals based on the inertial sensor with a blind source separation method;a matched-filter-based coherent integration method for passive bistatic radar using frequency-agile radar as the illuminator;research on recognition of CHD heart sound using MFCC and LPCC;and ultrasonic speckle tracking with an adaptive frame interval for the measurement of blood flow velocities.
Frequency invariant beamformers (FIBs) are widely used in speech enhancement and source localization. There are two traditional optimization methods for FIB design. The first one is convex optimization, which is simpl...
详细信息
Frequency invariant beamformers (FIBs) are widely used in speech enhancement and source localization. There are two traditional optimization methods for FIB design. The first one is convex optimization, which is simple but the frequency invariant characteristic of the beam pattern is poor with respect to frequency band of five octaves. The least squares (LS) approach using spatial response variation (SRV) constraint is another optimization method. Although, it can provide good frequency invariant property, it usually couldn't be used in speech enhancement for its lack of weight norm constraint which is related to the robustness of a beamformer. In this paper, a robust wideband beamforming method with a constant beamwidth is proposed. The frequency invariant beam pattern is achieved by resolving an optimization problem of the SRV constraint to cover speech frequency band. With the control of sidelobe level, it is available for the frequency invariant beamformer (FIB) to prevent distortion of interference from the undesirable direction. The approach is completed in time-domain by placing tapped delay lines(TDL) and finite impulse response (FIR) filter at the output of each sensor which is more convenient than the Frost processor. By invoking the weight norm constraint, the robustness of the beamformer is further improved against random errors. Experiment results show that the proposed method has a constant beamwidth and almost the same white noise gain as traditional delay-and-sum (DAS) beamformer.
暂无评论