Numerical differentiation of noisy data with local optimum by data segmentation

A new numerical differentiation method with local opti- mum by data segmentation is proposed. The segmentation of data is based on the second derivatives computed by a Fourier devel- opment method. A filtering process is used to achieve acceptable segmentation. Numerical results are presented by using the data segmentation method, compared with the regularization method. For further investigation, the proposed algorithm is applied to the resistance capacitance （RC） networks identification problem, and improvements of the result are obtained by using this algorithm.

A Partition Checkpoint Strategy Based on Data Segment Priority

A partition checkpoint strategy based on data segment priority is presented to meet the timing constraints of the data and the transaction in embedded real-time main memory database systems（ERTMMDBS） as well as to reduce the number of the transactions missing their deadlines and the recovery time.The partition checkpoint strategy takes into account the characteristics of the data and the transactions associated with it;moreover,it partitions the database according to the data segment priority and sets the corresponding checkpoint frequency to each partition for independent checkpoint operation.The simulation results show that the partition checkpoint strategy decreases the ratio of trans-actions missing their deadlines.

Individualization of Data-Segment-Related Parameters for Improvement of EEG Signal Classification in Brain-Computer Interface

In electroencephalogram (EEG) modeling techniques, data segment selection is the first and still an important step. The influence of a set of data-segment-related parameters on feature extraction and classification in an EEG-based brain-computer interface (BCI) was studied. An auto search algorithm was developed to study four datasegment-related parameters in each trial of 12 subjects’ EEG. The length of data segment (LDS), the start position of data (SPD) segment, AR order, and number of trials (NT) were used to build the model. The study showed that, compared with the classification ratio (CR) without parameter selection, the CR was increased by 20% to 30% with proper selection of these data-segment-related parameters, and the optimum parameter values were subject-dependent. This suggests that the data-segment-related parameters should be individualized when building models for BCI.

Relationship of Uncertainty Between Polygon Segment and Line Segment for Spatial Data in GIS

The mathematic theory for uncertainty model of line segment are summed up to achieve a general conception, and the line error band model of ε_σ is a basic uncertainty model that can depict the line accuracy and quality efficiently while the model of ε_m and error entropy can be regarded as the supplement of it. The error band model will reflect and describe the influence of line uncertainty on polygon uncertainty. Therefore, the statistical characteristic of the line error is studied deeply by analyzing the probability that the line error falls into a certain range. Moreover, the theory accordance is achieved in the selecting the error buffer for line feature and the error indicator. The relationship of the accuracy of area for a polygon with the error loop for a polygon boundary is deduced and computed.

Semantic Segmentation Based Remote Sensing Data Fusion on Crops Detection

Data fusion is usually an important process in multi-sensor remotely sensed imagery integration environments with the aim of enriching features lacking in the sensors involved in the fusion process. This technique has attracted much interest in many researches especially in the field of agriculture. On the other hand, deep learning (DL) based semantic segmentation shows high performance in remote sensing classification, and it requires large datasets in a supervised learning way. In the paper, a method of fusing multi-source remote sensing images with convolution neural networks (CNN) for semantic segmentation is proposed and applied to identify crops. Venezuelan Remote Sensing Satellite-2 (VRSS-2) and the high-resolution of Google Earth (GE) imageries have been used and more than 1000 sample sets have been collected for supervised learning process. The experiment results show that the crops extraction with an average overall accuracy more than 93% has been obtained, which demonstrates that data fusion combined with DL is highly feasible to crops extraction from satellite images and GE imagery, and it shows that deep learning techniques can serve as an invaluable tools for larger remote sensing data fusion frameworks, specifically for the applications in precision farming.

Research Model of Churn Prediction Based on Customer Segmentation and Misclassification Cost in the Context of Big Data

Enterprises have vast amounts of customer behavior data in the era of big data. How to take advantage of these data to evaluate custom forfeit risks effectively is a common issue faced by enterprises. Most of traditional customer churn predicting models ignore customer segmentation and misclassification cost, which reduces the rationality of model. Dealing with these deficiencies, we established a research model of customer churn based on customer segmentation and misclassification cost. We utilized this model to analyze customer behavior data of a telecom company. The results show that this model is better than those models without customer segmentation and misclassification cost in terms of the performance, accuracy and coverage of model.

Distributed C-Means Algorithm for Big Data Image Segmentation on a Massively Parallel and Distributed Virtual Machine Based on Cooperative Mobile Agents

The aim of this paper is to present a distributed algorithm for big data classification, and its application for Magnetic Resonance Images (MRI) segmentation. We choose the well-known classification method which is the c-means method. The proposed method is introduced in order to perform a cognitive program which is assigned to be implemented on a parallel and distributed machine based on mobile agents. The main idea of the proposed algorithm is to execute the c-means classification procedure by the Mobile Classification Agents (Team Workers) on different nodes on their data at the same time and provide the results to their Mobile Host Agent (Team Leader) which computes the global results and orchestrates the classification until the convergence condition is achieved and the output segmented images will be provided from the Mobile Classification Agents. The data in our case are the big data MRI image of size (m × n) which is splitted into (m × n) elementary images one per mobile classification agent to perform the classification procedure. The experimental results show that the use of the distributed architecture improves significantly the big data segmentation efficiency.

Analysis of an Active Fault Geometry Using Satellite Sensor and DEM Data: Gaziköy-Saros Segment (NAFZ), Turkey

In this study, Landsat 5 Thematic Mapper (TM) and SPOT HRV Panchromatic data were analysed to determine the geometry of an active fault segment (the Ganos segment) in Gazikoy-Saros region, west of Marmara Sea, Turkey. Gazikoy-Saros/Ganos segment is a part of North Anatolian Fault Zone (NAFZ). North-Anatolian fault is considered to be one of the most important active strike-slip faults in the world. Thus far in relevant researches based on Gazikoy-Saros segment a single straight fault line representation is used on the fault descriptive geological maps. This study, with the aid of enhanced remotely sensed data aims to reveal the linear details of the NAFZ fault segment, which subsequently were superposed with a Digital Elevation Model (DEM) data. Respectively, using these data the surface geometry expression of Gazikoy-Saros fault segment was detailed and remapped. According to the results of the analysis two small releasing steps were identified on this segment. The first one is situated between Mürseli and Güzelkoy villages, and the second one is between Mürseli and Yorguc villages. In addition to this, it is found that the fault strike bends approximately 7° further to in south-eastern (SE) direction between Yenikoy and Sofular villages. This angular change was defined with the advantage of multi-angular viewing capability of the multi-satellite sensors and DEM data. The newly generated surface geometry expression of Ganos segment was compared with Global Positioning System (GPS) velocity vectors.

Interpretation of the west segment of the coastal fault zone in the coastal region of South China based on the gravity data

By systemic processing, comprehensive analysis, and interpretation of gravity data, we confirmed the existence of the west segment of the coastal fault zone(west of Yangjiang to Beibu Bay) in the coastal region of South China. This showed an apparent high gravity gradient in the NEE direction, and worse linearity and less compactness than that in the Pearl River month. This also revealed a relatively large curvature and a complicated gravity structure. In the finding images processed by the gravity data system, each fault was well reflected and primarily characterized by isolines or thick black stripes with a cutting depth greater than 30 km. Though mutually cut by NW-trending and NE-trending faults, the apparent NEE stripe-shaped structure of the west segment of the coastal fault zone remained unchanged,with good continuity and an activity strength higher than that of NW and NE-trending faults. Moreover,we determined that the west segment of the coastal fault zone is the major seismogenic structure responsible for strong earthquakes in the coastal region in the border area of Guangdong, Guangxi, and Hainan.

基于边界点估计与稀疏卷积神经网络的三维点云语义分割

针对大规模点云具有稀疏性,传统点云方法提取上下文语义特征不够丰富,并且语义分割结果存在物体边界模糊的问题,提出基于边界点估计与稀疏卷积神经网络的三维点云语义分割算法,主要包括体素分支与点分支.对于体素分支,将原始点云进行体素化后经过稀疏卷积得到上下文语义特征;进行解体素化得到每个点的初始语义标签;将初始语义标签输入到边界点估计模块中得到可能的边界点.对于点分支,使用改进的动态图卷积模块提取点云局部几何特征;依次经过空间注意力模块与通道注意力模块增强局部特征;将点分支得到的局部几何特征与体素分支得到的上下文特征融合,增强点云特征的丰富性.本算法在S3DIS数据集和SemanticKITTI数据集上的语义分割精度分别达到69.5%和62.7%.实验结果表明,本研究算法能够提取到更丰富的点云特征,可以对物体的边界区域进行准确分割,具有较好的三维点云语义分割能力.

改进Hybrid-Task-Cascade的染色体分割研究

针对人工分割染色体图像中的实例存在耗时长,精度不佳等问题,提出一种基于改进Hybrid Task Cascade模型的染色体实例分割方法。首先,提出基于实例操作的染色体增强策略,以扩充量少且信息不丰富的染色体数据集。然后使用PAFPN代替Hybrid Task Cascade中的特征金字塔模块,保留浅层特征信息,提高定位和分割染色体实例的准确度。针对重叠的染色体簇引入Soft-NMS方法改进候选框的筛选,保留更多的染色体包围框。最后,将测试集的结果与其它模型进行对比,采用平均准确率(mean average precision, mAP)、AP50和AP75评估模型定位和分割性能,通过对自采集的染色体图像进行评估验证,其包围框定位平均准确率和分割平均准确率分别达到了80.53%和77.55%。实验表明,上述方法在染色体图像数据集上具有较好的分割效果。

改进自适应加权的海面目标距离测量和跟踪

海上目标距离探测和跟踪是海洋安全和军事应用中的关键任务之一,针对复杂海面环境在海杂波等影响因素下目标距离测量精度低的问题,提出了一种改进的多传感器数据融合算法。该算法利用海面舰船雷达及陆地雷达联合探测结果,先对多个数据源数据进行坐标系转换,利用Robust Z-score方法进行纵向数据预处理剔除异常数据,再通过重新定义置信距离度量,将置信度较高的传感器结果代替被踢除数据后,对结果进行自适应加权融合。同时,为了进一步提高数据精度,引入了一种分段融合机制,将改进的传感器数据融合算法与阶梯式自适应加权融合算法进行级联,通过度量各分段融合结果的相似度,设定一个置信阈值,通过该置信阈值确定最终的融合结果。仿真实验结果证实了算法的有效性和准确性。

DNA测序首段信号数据特别处理方法研究

针对荧光标记毛细管电泳DNA测序开始一段序列的信号杂乱特性以及不同碱基通道信号间的相对偏移问题,提出了一种用于首段信号特别处理的电泳迁移率自适应校正方法。分析设计基于四通道首段信号间峰重叠程度估计的运筹决策目标函数,通过目标函数动态寻优计算确定不同碱基通道信号的相对偏移系数,从而实现DNA测序首段信号的相对偏移校正,为提高首段信号碱基的合理判读及后续DNA精准排序提供了可靠的过程数据。

改进的3D-BoNet算法应用于点云实例分割与三维重建

为了更好地利用点云数据重建室内三维模型,本文提出了一种基于3D-BoNet-IAM算法的室内场景三维重建方法。该方法通过改进3D-BoNet算法提高点云数据的实例分割精度。针对点云数据缺失问题,提出了基于平面基元合并优化的拟合平面方法,利用拟合得到的新平面重建建筑表面模型。在S3DIS和ScanNet V2数据集上验证3D-BoNet算法的改进效果。试验结果表明,本文提出的3D-BoNet-IAM算法比原始算法分割精度提高了3.3%;对比本文建模效果与其他建模效果发现,本文方法的建模效果更准确。本文方法能够提高室内点云数据的实例分割精度,同时得到高质量的室内三维模型。

基于多源数据的街道空间品质测度研究——以芜湖市中心城区为例

在20世纪80年代后,城市经济与高速公路的发展使城市结构发生剧变,由街道构成的传统城镇空间形态被打破。而街道空间作为城市公共空间的重要组成部分,其空间品质的研究对城市在打造魅力街道、传统特色保留以及时代新元素的融入方面具有重大意义。研究发现:芜湖市中心城区街道综合空间品质整体上,呈现出中心放射状的整体结构,城市空间品质测度结构及城市建设强度的重心数值也呈现出南高北低、内高外低的指状分布特征;芜湖市中心城区5种类型的街道在空间分布上表现出较为分散的特征,不同类型的街道聚类伴随区位的迁移具有明显的差异性;交通导向型街道趋向于城市干道及快速路,但由于城市的各类服务型业态难以覆盖而导致街道服务性不足,生活导向型街道多数位于城市核心建设区,需要加强街道绿化和空间开敞度方面的建设。

基于深度学习的移动机器人语义SLAM方法研究

为了给移动机器人提供细节丰富的三维语义地图,支撑机器人的精准定位,本文提出一种结合RGB-D信息与深度学习结果的机器人语义同步定位与建图方法。改进了ORB-SLAM2算法的框架,提出一种可以构建稠密点云地图的视觉同步定位与建图系统;将深度学习的目标检测算法YOLO v5与视觉同步定位与建图系统融合,反映射为三维点云语义标签,结合点云分割完成数据关联和物体模型更新,并用八叉树的地图形式存储地图信息;基于移动机器人平台,在实验室环境下开展移动机器人三维语义同步定位与建图实验,实验结果验证了本文语义同步定位与建图算法的语义信息映射、点云分割与语义信息匹配以及三维语义地图构建的有效性。

基于特征与数据增强的城市街景实例分割算法

城市街景分割是智能交通领域中一项关键的技术,对于城市街景环境中的客观因素例如遮挡、小目标等问题,提出一种基于特征增强与数据增强的城市街景实例分割算法DF-SOLO(data augmentation and feature en-hancement SOLO)。针对遮挡问题,通过非对称自编-解码器架构对城市街景图像进行数据增强,与传统方法相比处理后的图像更贴近真实的源数据分布。针对城市街景中的小目标分割问题,引入特征加权和特征融合的思想,特征加权模块在特征处理过程中能够根据特征的重要程度赋予不同的权值,提高对重要特征的利用率;特征融合模块从更细粒度的角度进行多尺度特征融合以解决尺度敏感问题,提高语义特征的描述性。通过在Cityscapes数据集上的实验表明,提出的实例分割算法在保证实时性的同时相较于单阶段SOLO算法和两阶段Mask R-CNN算法的mAP值上分别提升2.1%和2%,改善了对小目标和遮挡目标的分割效果。

基于Transformer的城市三角网格语义分割方法

对城市三角网格(Urban Triangle Mesh)数据进行语义分割以识别不同类别的物体,是理解和分析三维城市场景的一种非常重要的方法.城市三角网格是一种具有丰富空间拓扑关系的三维空间几何数据,包含大量的几何信息,然而,现有的方法仅仅单独对每种几何信息进行特征提取,然后简单地融合再进行语义分割,难以利用几何信息之间的关联性,对个别物体的分割性能不佳.为了解决上述问题,提出一种基于自注意力机制Transformer的模型UMeT(Urban Mesh Transformer),其由多层感知机和MeshiT(Mesh in Transformer)模块构成,不仅可以利用多层感知机提取高维特征,还可以利用MeshiT模块计算各种几何信息之间的关联性,有效挖掘城市三角网格数据中隐含的关联.实验证明,UMeT能提取高维特征,同时保证城市三角网格数据的空间不变性,从而提升了语义分割的准确性.

一种工控协议识别中的特征字符串挖掘算法

对工控协议的识别,是对工控协议开展研究的第一步。而在通信过程中频繁出现的字符串,是对工控协议识别中的重要特征。针对工控协议识别中特征字符串的提取问题,提出了一种自顶向下的频繁字符串挖掘算法,可以直接得到没有冗余的频繁字符串集。同时,对于自顶向下方法中原始数据过于庞大、算法迭代次数较多等问题,借鉴了N-gram模型,提出了一种数据划分策略,解决了自顶向下处理时数据过大的问题。此外,在挖掘频繁字符串的过程中,采取了删除重叠项与字符串分裂相结合的方法。实验结果表明,该算法针对多种协议均能识别出其中的特征字符串;同时,利用识别出的字符串作为特征,在协议识别工作中也能取得良好的效果。可以得出结论,该算法能够较好地提取出工控协议中的特征字符串。