Although SISR (Single Image Super Resolution) problem can be effectively solved by deep learning based methods, the training phase often considers single degradation type such as bicubic interpolation or Gaussian blur with fixed variance. These priori hypotheses often fail and lead to reconstruction error in real scenario. In this paper, we propose an end-to-end CNN model RPSRMD to handle SR problem in multiple Gaussian degradations by extracting and using as side information a shared image prior that is consistent in different Gaussian degradations. The shared image prior is generated by an AED network RPGen with a rationally designed loss function that contains two parts: consistency loss and validity loss. These losses supervise the training of AED to guarantee that the image priors of one image with different Gaussian blurs to be very similar. Afterwards we carefully designed a SR network, which is termed as PResNet (Prior based Residual Network) in this paper, to efficiently use the image priors and generate high quality and robust SR images when unknown Gaussian blur is presented. When we applied variant Gaussian blurs to the low resolution images, the experiments prove that our proposed RPSRMD, which includes RPGen and PResNet as two core components, is superior to many state-of-the-art SR methods that were designed and trained to handle multi-degradation.
With the rapid advancement of drone technology, there has been a significant increase in the demand for detecting “Low, Slow, and Small” (LSS) drones. However, existing deep learning networks often overlook critical details of such small targets during feature extraction, leading to reduced detection accuracy. Furthermore, edge inference devices typically have limited computational power, which hampers real-time processing capabilities. To address these challenges, we propose a lightweight and efficient drone detection network called NFE-YOLO. This network introduces an efficient positive traffic channel attention module known as EOrthoNet and enhances small target detection capabilities by improving the model’s neck structure. Specifically, the neck incorporates partial Convolution and C3Faster modules to reduce model size while the EOrthoNet channel attention module improves feature extraction accuracy.To further enhance the model’s generalization ability, we present a cross-scenario visible-light “Low, Slow, and Small” drone dataset referred to as UESTC Anti-UAV. This dataset comprises a total of 10,099 images that have all been manually annotated with precision to address the issue of insufficient sample diversity in existing datasets.Experimental results demonstrate a significant improvement in detection accuracy across multiple datasets. NFE-YOLO achieves a mAP50 value of 0.987 on the UESTC Anti-UAV dataset representing a 2.3% improvement over YOLOv8n with a model size of only 3.72MB. The proposed method effectively enhances detection accuracy for LSS drones and is crucial for the development of lightweight drone detection models.
28 January 2014 in EURASIP Journal on Wireless Communications and Networking
A critical objective of cognitive radio (CR) system is to enhance the spectrum efficiency, and one of the key factors that can determine the spectrum efficiency is the system spectrum sensing performance with respect to sensing window length. For non-coherent detection-based sensing technique, the length of the sensing window required to meet the detection criteria is inversely proportional to the detected signal-to-noise ratio (SNR) of the primary users (PUs). This fact may result in an inadequate use of the white or gray space for the conventional CR system whose transmission and sensing window length are both fixed because a high detected PUs SNR can lead to an excessive long fixed sensing window which occupies the potential CR transmission opportunities while a low received PUs SNR can result in an insufficient sensing window length which degrades the CR detection criteria. In this paper, to improve the spectrum efficiency compared with the fixed sensing/transmission window length-based CR system, we propose an adaptive spectrum sensing window length optimization algorithm. We design the algorithm based on the ultra wideband (UWB) system which is an ideal candidate for the implementation of the CR technology. Based on the analysis of the CR-UWB’s spectrum sensing technique in terms of the factors such as spectrum efficiency, spectrum sensing length, PUs SNR, detection criteria etc., we formulate the optimization problem into a convex problem, which enables the proposed algorithm to find the optimal trade-off with low computational complexity between the sensing window length and the desired detection probabilities for the CR-UWB system. Compared with the conventional fixed length spectrum sensing techniques, the proposed algorithm is verified to be able to adapt the length of the CR-UWB’s transmission window according to the PUs SNR to optimize the use of the available spectrum while guaranteeing the PUs from being interfered.
Ultra Wideband (UWB) system uses extremely low signal power to coexist with other primary users (PUs), which limits the use of the UWB's wide spectrum band. Cognitive Radio (CR) enables UWB systems to efficiently use the overlapped spectrum without causing interference to other wireless systems. In this paper, we focuses on the low-complexity joint optimization algorithm design with respect to transmit power allocation and spectrum sensing time (SST) for maximizing the spectrum efficiency of the Orthogonal Frequency Division Multiplexing based CR-UWB system. The SST optimization algorithm minimizes the spectrum sensing time in order to maximize the time length of applying the power allocation algorithm for data transmission. The proposed group power allocation algorithm adaptively assigns the transmit power to the subcarrier groups according to the effective signal-to-noise ratio (SNR) of each subcarrier group based on greedy algorithm. The proposed joint optimization algorithm can maximize the CR-UWB systems spectrum efficiency at a extremely low primary user SNR regime with low complexity.
As an alternative to the traditional pilot-aided orthogonal frequency division multiplexing (OFDM), the time-domain pseudonoise (PN)-padded OFDM provides a higher spectral efficiency. However, the carrier frequency offset (CFO) attenuates peaks of the conventional PN correlation output, which limits the CFO estimation range of the OFDM synchronizer. An improved correlation is proposed in this letter to remove the CFO-induced amplitude attenuation of correlation peaks. For a synchronizer adopting the designed correlator, a larger range of CFO acquisition is obtained through using wider correlation windows with a smaller interval between them. The proposed method of CFO acquisition is verified in a digital terrestrial multimedia broadcast receiver, in which the synchronizer is able to acquire CFOs up to ±320 kHz in the DVB-T F1 channel. Furthermore, the acquisition range can be expanded in more favorable channels.