Subjects: Electronics and Communication Technology >> Information Processing submitted time 2024-01-07
Abstract： In recent years, with the rapid development of deep learning technology, image recognition based on convolutional neural networks (CNN) has achieved remarkable achievements in various fields. In the field of botany, flower species identification is an important research direction and is of great significance in ecology, agriculture, and environmental monitoring. This research aims to explore and optimize the application of ResNet (deep residual network) in flower type recognition tasks. First, the article conducts an in-depth analysis of the structure of the ResNet network, understands its mechanism for introducing residual learning, and how to effectively deal with the vanishing and exploding gradient problems in deep network training. Through preliminary experiments on a large-scale flower image data set, the excellent performance of ResNet in handling complex multi-category flower image recognition tasks was verified. In the data preprocessing stage, the article uses data enhancement techniques, including cropping and flipping, to expand the training data set and improve the generalization ability of the model. At the same time, the flower images are standardized to adapt to the requirements of the ResNet network for input data. Experimental results show that compared with the traditional CNN model, the flower type recognition model using ResNet has significantly improved accuracy and convergence speed. In addition, through in-depth analysis of the model's performance on different flower categories, it was found that the ResNet network performed better when processing flower images with hierarchical structures and complex shapes. The model proposed in this article not only achieves excellent performance overall, but also has high accuracy in identifying specific flower categories. In further research, we consider further improving the generalization ability of the model through transfer learning, especially when facing small sample flower data sets. At the same time, the real-time performance of the model will be explored to adapt to the need for rapid and accurate identification of flower types in real scenes.
This study provides a useful reference and reference for the application of deep learning in the field of botany by conducting a comprehensive and in-depth analysis of the advantages and applications of ResNet network in flower species recognition tasks. The research results not only have certain theoretical value for the improvement of flower identification technology, but also have extensive potential for promotion in practical applications.
Peer Review Status:
Awaiting Review
Subjects: Electronics and Communication Technology >> Information Processing submitted time 2023-11-07 Cooperative journals: 《文献与数据学报》
Abstract： ［Purpose/significance］Data fusion is an important way to realize multi-source data value.Comprehensive analysis of the overall topics of global data fusion research has an important scientific and technological information value for the current data fusion research. ［Method/process］The hot topics and research methods of 16053 literatures from Web of Science core collections were analyzed by word-frequency and co-word analysis. ［Result/conclusion］The data fusion research has shown a significant growth trend, and after more than 30 years development, core research hotspots and methods of data fusion have been formed. In the research, the data fusion of sensors （including wireless sensors） is the core research direction in this field. Fault diagnosis, remote sensing, security and smart grid are the hotspots of the data fusion scenario. Kalman Filter, Neural Network, Dempster-Shafer Evidence Theory and Machine Learning（including Deep Learning, Support Vector Machine, etc.） are the main methods in data fusion, and the synergy network of methods have been formed in data fusion.
Subjects: Electronics and Communication Technology >> Information Processing submitted time 2022-06-07 Cooperative journals: 《桂林电子科技大学学报》
Abstract：在无线传感器网络（WSNs）中，数据分发是一种重要的传输方式，需要满足以下要求：可靠性，节能和可扩展性。然而，现有的研究工作很少关注数据分发中所存在的攻击，导致数据分发的可靠性大打折扣。为了检测出WSNs数据分发中的节点克隆攻击，保证数据分发的高可靠性，提出了基于单轮零知识证明的节点克隆攻击检测方案。本方案通过构建析取-叠加码生成专属于各个节点的数字指纹，在单轮零知识证明方案中对节点的数字指纹进行验证，可以检测出没有正确数字指纹的克隆节点。仿真表明使用提出的检测方案，可以保证WSNs在数据分发过程中的高可靠性. In wireless sensor networks (WSNs), data dissemination is an essential transmission mode, which needs to meet three requirements: reliability, energy saving and scalability. Existing research pays little attention to the attacks in data dissemination, resulting in a significant loss of the reliability of data dissemination. In order to detect node clone attacks in WSNs data dissemination and ensure high reliability of data dissemination, a node clone attack detection scheme based on single round zero knowledge proof is proposed. In this scheme, the digital fingerprint of each node is generated by constructing superimposed disjunct code, and the cloned node without correct digital fingerprint can be detected by verifying the digital fingerprint of the node in the single round zero-knowledge proof scheme. Simulation results show that the proposed scheme can ensure the high reliability of WSNs in the data dissemination process.
Subjects: Electronics and Communication Technology >> Information Processing Subjects: Engineering and technical science >> Engineering General Technology Subjects: Mechanical Engineering >> Other Disciplines of Mechanical Engineering submitted time 2022-02-08
Abstract：The spectral leakage (SL) from windowing and the picket fence effect (PEF) from discretization have been among the standard contents in textbooks for many decades. The SL and PEF would cause the distortions in amplitude, frequency, and phase of signals, which have always been of concern, and attempts have been made to solve them. This paper proposes two novel decomposition theorems that can totally eliminate the SL and PEF, they could broaden the knowledge of signal processing. First, two generalized eigenvalue equations are constructed for multifrequency discrete real signals and complex signals. The two decomposition theorems are then proved. On these bases, exact decomposition methods for real and complex signals are proposed. For a noise-free multifrequency real signal with m sinusoidal components, the frequency, amplitude, and phase of each component can be exactly calculated by using just 4m1 discrete values and its second-order derivatives. For a multifrequency complex signal, only 2m1 discrete values and its first-order derivatives are needed. The numerical experiments show that the proposed methods have very high resolution, and the sampling rate does not necessarily obey the Nyquist sampling theorem. With noisy signals, the proposed methods have extraordinary accuracy.
Peer Review Status:Awaiting Review
Subjects: Electronics and Communication Technology >> Information Processing submitted time 2021-09-10
Abstract： The geostationary optical remote sensing satellite has the advantages of high temporal resolution and wide coverage, which can continuously track and observe ship targets on the sea in a large range. However, the ship targets in geostationary satellite remote sensing image are usually small and weak, and are easily affected by cloud, island and other factors, which brings great difficulty to the detection of ship targets. This paper proposes a new method for detecting ships moving on the sea surface from geostationary optical remote sensing images: Firstly, the adaptive nonlinear gray stretch (ANGS) method is used to enhance the image to highlight the small and weak ship targets. Secondly, a multi-scale dual-neighbor difference contrast measure (MDDCM) method is designed to detect the position of the candidate ship target. Then, the shape characteristics of each candidate area is analyzed to remove false ship targets. Finally, the joint probability data association (JPDA) method is used for multi-frame data association and tracking. Experiments show that the proposed method can effectively detect and track moving ship targets in GF-4 satellite optical remote sensing images, and the method has better detection performance compared with other classical methods. "
Peer Review Status:Awaiting Review
Subjects: Electronics and Communication Technology >> Information Processing submitted time 2018-08-31
Abstract： " This paper innovatively proposes a technical idea that uses image deep learning to solve the problem of radio signal recognition: first, it transforms the radio signal into a two-dimensional picture, and transforms the radio signal recognition problem into the object detection problem in the field of image recognition; then, it makes use of the advanced achievements about image recognition to improve the intelligence and ability of radio signal recognition in complex electromagnetic environment. Based on this idea, a novel radio signal recognition algorithm named RadioImageDet is proposed in this paper. The experimental results show that the algorithm can effectively identify the waveform types and time/frequency coordinates of radio signals. After training and testing on the self-collected data set with 12 types and 4740 samples, the accuracy reaches 86.03% and the mAP value reaches 77.72, while the detection time is only 33 milliseconds on the medium configured desktop computer. "
Peer Review Status:Awaiting Review
Subjects: Electronics and Communication Technology >> Information Processing submitted time 2017-04-19
Abstract：摘要：传统子空间类方法依赖于阵列相关矩阵，在阵列中阵元出现失效的情况下，相关矩阵将会秩亏，子空间类方法将会失效。针对该问题，本文从协方差拟合准则出发，将协方差矩阵的 Toeplitz 结构作为约束，基于低秩矩阵重构的原理来恢复协方差矩阵，而后采用子空间类方法来进行目标的方位估计。在阵元失效情况下，该方法能有效重构阵列协方差矩阵，恢复失效阵元的自由度，解决失效阵元情况下高精度目标方位估计的问题。数值仿真表明，该方法在阵元失效条件下，能够恢复损伤阵列到正常阵列条件下的性能，尤其是对于多个目标的情况，该方法表现出更优的性能。
Peer Review Status:Awaiting Review
Hits
Hits
Downloads
Comment
YES