检索结果-南通市图书馆

同方期刊数据库博看期刊评论

在线全文

学校读者我要写书评

暂无评论

基于深度卷积判别网络的人脸比对方法

哈尔滨工程大学学报 2024年第9期45卷 1770-1782页

作者：谷凤伟陆军刘子玄蔡成涛哈尔滨工程大学智能科学与工程学院黑龙江哈尔滨150001 哈尔滨工程大学船海装备智能化技术与应用教育部重点实验室黑龙江哈尔滨150001

针对实际应用中人脸比对面临着场景复杂性高、光照、遮挡等问题,为了提高人脸比对准确率,本文提出了一种基于深度卷积判别网络的人脸比对算法MTC-FaceNetSDM。建立了MTC-FaceNetSDM的深度卷积神经网络,在FaceNet网络前端中融合多任务级... 详细信息

针对实际应用中人脸比对面临着场景复杂性高、光照、遮挡等问题,为了提高人脸比对准确率,本文提出了一种基于深度卷积判别网络的人脸比对算法MTC-FaceNetSDM。建立了MTC-FaceNetSDM的深度卷积神经网络,在FaceNet网络前端中融合多任务级联卷积神经网络得到MTC-FaceNet网络,实现实际场景中的人脸检测提取目标人脸;利用深度卷积神经网络获取高维人脸深度特征,并将FaceNet网络的欧氏距离模块替换为所提出的相似度判别模块SDM,用于高维人脸特征向量比对;最终,利用自制的人脸数据集C-facev1,结合CASIA-WebFace人脸数据集对本文人脸比对算法进行训练,使用人脸数据集LFW和CASIA-FaceV5对训练后的模型进行性能评估。实验结果表明:本文所设计的MTC-FaceNetSDM的人脸比对准确率比MTC-FaceNet整体提高1.48%,对中国人脸比对准确率提高3.80%,可实现多人种的人脸比对,同时该算法具备良好的鲁棒性和泛化能力,达到优良的人脸比对效果,可实际应用于人脸验证系统。

关键词：人脸比对深度卷积判别网络多任务级联卷积神经网络相似度判别模块人脸特征向量

Helicopter blades running elevation measurement using omnidirectional vision

同方期刊数据库博看期刊评论

在线全文

学校读者我要写书评

暂无评论

Chinese Journal of Aeronautics 2017年第6期30卷 1974-1984页

作者： chengtao cai Xiangyu WENG College of Automation Harbin Engineering University Harbin 150001 China

Omnidirectional dynamic space parameters of high-speed rotating helicopter blades are precise 3 D vector description of the blades. In particular, the elevation difference is directly related to the aerodynamic performance and maneuverability of the helicopter. The state of the art detection techniques based on optics and common vision have several drawbacks, such as high demands on devices but poor extensibility, limited measurement range and fixed measurement position. In this paper, a novel approach of helicopter blades running elevation measurement is proposed based on omnidirectional vision. With the advantages of panoramic visual imaging integration, 360° field of view and rotation in-variance, high-resolution images of all rotating blades positions are obtained at one time. By studying the non-linear calibration and calculation model of omnidirectional vision system, aiming at solving the problem of inaccurate visual space mapping model,the omnidirectional and full-scale measurement of the elevation difference are finalized. Experiments are carried out on our multifunctional simulation blades test system and the practical blades test tower, respectively. The experimental results demonstrate the effectiveness of the proposed method and show that the proposed method can considerably reduce the complexity of measurement.

关键词： Full-scale measurement Helicopter blades elevation Non-linear calibration Omnidirectional vision Unified sphere model

基于加幂积分法的水面船舶有限时间轨迹跟踪控制

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

船舶 2024年第2期35卷 9-18页

作者：曾道辉蔡成涛哈尔滨工程大学智能科学与工程学院哈尔滨150001 船海装备智能化技术与应用教育部重点实验室哈尔滨150001 电子政务建模仿真国家工程实验室哈尔滨150001

针对具有在模型参数不确定性和未知时变扰动下的水面船舶,该文提出1种基于扩张状态观测器的有限时间轨迹跟踪控制方案。首先,基于反步法和加幂积分法,提出了1种新的有限时间控制律,所设计的控制器通过采用连续反馈,可确保水面船舶的轨... 详细信息

针对具有在模型参数不确定性和未知时变扰动下的水面船舶,该文提出1种基于扩张状态观测器的有限时间轨迹跟踪控制方案。首先,基于反步法和加幂积分法,提出了1种新的有限时间控制律,所设计的控制器通过采用连续反馈,可确保水面船舶的轨迹跟踪误差在有限时间内收敛到原点附近的邻域内;其次,在李雅普诺夫意义下严格证明了闭环控制系统的有限时间稳定性;最后,通过仿真比较表明了所提出的控制方案的优越性。

关键词：水面船舶轨迹跟踪扩张状态观测器有限时间控制器加幂积分法

The Design of Aviation Battery Charging and Discharging System based on DSP

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

The Design of Aviation Battery Charging and Discharging Syst...

第26届中国控制与决策会议

作者： cai chengtao YANG Chao LV Xiaolong College of Automation Harbin Engineering University

ISBN: (纸本)9781479937066

Aviation battery which is used for giving power to critical load of aircraft for an emergency situation is an important part of the aircraft airborne equipment Traditional battery charging equipments usually adopt simple methods such as constant current or constant voltage to charge battery without the ability of monitoring the status of battery in real time in the charging and discharging process. This imprudent approach usually leads to the over-charging as well as irreversible damage to the battery. For dealing with challenge problem, we intend to apply modern power electronics technology, high-frequency inverter technology and discharge depolarization timely technology to overcome those shortcomings mentioned above. At the same time, our major task focused on the hardware design based on DSP controller, including AC to DC circuit, full bridge inverter circuit, driver circuit, the rectifier filter circuit and so on. The simulation and experimental results show that the aviation battery charging equipment we designed can fulfill the practical requirement and has more advantages such as high effective and real-time status detection etc.

关键词： Aviation battery intelligent charging and discharging charging equipment full bridge inverter

新工科背景下“五融入”创新人才培养模式研究——以“机器学习”课程为例

在线全文

cnki会议

学校读者我要写书评

暂无评论

科教导刊 2024年第1期 80-82页

作者：王兴梅杨东梅蔡成涛哈尔滨工程大学计算机科学学院黑龙江哈尔滨150001

机器学习作为培养人工智能领域人才的核心课程,其在教学内容、教学方式、课程考核等环节可进一步研究与改进。基于新工科背景,为培养顺应时代发展、符合企业需求的人才,文章以“机器学习”课程为例,探索了“五融入”创新人才培养模式,... 详细信息

机器学习作为培养人工智能领域人才的核心课程,其在教学内容、教学方式、课程考核等环节可进一步研究与改进。基于新工科背景,为培养顺应时代发展、符合企业需求的人才,文章以“机器学习”课程为例,探索了“五融入”创新人才培养模式,提出将科学精神、科研成果、科研实践、企业资源和SPOC微课堂融入机器学习课程,着力推进知识传授、能力培养、思想引领三位一体的育人功能,也为我国开设同类课程的高校提供有益借鉴与参考。

关键词：机器学习人才培养课程思政校企合作

A Rapid and Precise Self-localization Approach of Mobile Robot Based on Binocular Omni-directional Vision

同方期刊数据库博看期刊评论

在线全文

学校读者我要写书评

暂无评论

A Rapid and Precise Self-localization Approach of Mobile Rob...

第36届中国控制会议

作者： Zhu Qidan Xie Hongle cai chengtao Liu Peng College of Automation Harbin Engineering University

ISBN: (纸本)9781538629185

In order to solve the problem of mobile robot self-localization,a robot vision absolute localization method based on binocular stereo omni-directional vision system is *** the artificial landmarks as a priori position information,the improved modified discrete gaussian hermite moment SURF algorithm is used to complete the rapid detection of feature points,as well as the feature matching algorithm to identify the area of the *** the matching center of gravity iteration algorithm,the matching error is *** algorithm of sub-pixel location of circle target based on unmixing the mixed pixels is applied to improve the positioning accuracy of the center coordinates of *** landmark detection area is predicted by the current state of motion of the robot,which improves the real-time positioning of robot while *** the characteristics of the panoramic system,a triangulation algorithm is improved to solve the localization and orientation of the *** to multiple localization experiments performed in different environmental paths,the results show that the proposed algorithm has high accuracy and real-time performance.

关键词： Binocular stereo localization Mobile robot Omni-directional vision SURF algorithm

Fusion of Infrared and Visible Image Based on HIS and Wavelet Transform

在线全文

cnki会议

学校读者我要写书评

暂无评论

Fusion of Infrared and Visible Image Based on HIS and Wavele...

第30届中国控制与决策会议

作者： chengtao cai Xin Ding Harbin Engineering University

Due to the different imaging principle of infrared image and visible image, the information characteristics of the two images are different. Fusion technology is the organic fusion of images from two different sensors, with complementary information. In this paper, a novel algorithm is proposed for the fusion of infrared and visible images. The algorithm are firstly performs HIS transform on the infrared image and the visible image to obtain two sets of intensity（I）, hue（H） and saturation（S） components. Then, after the wavelet transform of the I-components, a simple fusion algorithm based on spatial domain is adopted for low frequency sub-band, and the high-frequency sub-bands adopt the fusion algorithm based on window-based average gradient. Finally, the fused I-component was wavelet reconstructed and transformed by the inverse HIS transform with S-component and H-component of visible image to obtain a fused image. Experiments show that the novel fusion algorithm of infrared and visible image proposed in this paper can preserve the background information and texture information of the visible image better when the object is highlighted in the infrared image, the fused image is clearer, and the fusion effect is better.

关键词： Infrared and Visible Images HIS Transform Wavelet Transform Image Fusion

CNN-SELF-ATTENTION-DNN ARCHITECTURE FOR MANDARIN RECOGNITION

在线全文

cnki会议

学校读者我要写书评

暂无评论

CNN-SELF-ATTENTION-DNN ARCHITECTURE FOR MANDARIN RECOGNITION

第32届中国控制与决策会议

作者： chengtao cai Dongning Guo Harbin Engineering University

Connectionist temporal classification（CTC） is a frequently used approach for end-to-end speech *** can be used to calculate CTC loss with artificial neural network such as recurrent neural network（RNN） and convolutional neural network（CNN）.Recently,the self-attention architecture has been proposed to replace RNN due to its parallelism in *** this paper,we propose CNN-SELF-ATTENTION-DNN CTC architecture which use selfattention to replace RNN and combine with CNN and deep neural network（DNN）.We evaluate it with an about 170 hours Mandarin speech dataset and achieved 7% absolute character error rate（CER） reduction compared to DEEP-CNN with less training time than ***,we do an analysis on the test results to find the factors that affect CER.

关键词： Self-attention end-to-end Mandarin recognition CTC

The Application of the Dilated Convolution based on Small Object Detection

在线全文

cnki会议

学校读者我要写书评

暂无评论

The Application of the Dilated Convolution based on Small Ob...

第三十九届中国控制会议

作者： chengtao cai Yue Wu Shuofeng Li The college of Automation Harbin Engineering University

Object detection is an important part of computer vision applications. It is widely used in robot navigation, intelligent video surveillance, industrial testing, aerospace and other fields. Full application of computer vision can reduce the consumption of human capital. At present, most deep learning-based methods focus on large objects, and we know that large objects always account a large part of an image. Therefore, these methods are not ideal for detecting small objects. The reason is that small objects carry very little information and their ability to express features are also weak. To solve this problem, we use our model and try to enlarge the resolution of feature maps by applying the dilated convolution. The results show that it improves the accuracy of detecting small objects.

关键词： Object detection Small objects Dilated convolution