共查询到20条相似文献，搜索用时 31 毫秒
1.
Identification of forest fire-points in NOAA images in the basis of monitoring forest fire using NOAA satellite data. Traditional visual interpretation is difficult to settle for auto identification with computer. The artificial neural network technique provides a new means for solving this problem. In this paper, the principles and method of using neural network technique to automatically identify fire-points in NOAA images are discussed and the test in the range of Hubei province is presented. The result of the test shows that the disciplined neural network has collected the character of fire-points and has ability to identify fire-points in NOAA images. Comparing neural network with visual interpretation, the conclusion is drawn that by using neural network the purpose of auto-identification of forest fire-points in NOAA images can be realized with the almost same precision. 相似文献
2.
3.
基于高维云模型和RBF神经网络的遥感影像不确定性分类方法 总被引：1，自引：1，他引：0
云模型是用自然语言值表示的某个定性概念与其定量表示之间的不确定性转换模型,RBF神经网络已经广泛应用于遥感影像分类。考虑到传统的RBF神经网络分类技术不能有效表达影像分类过程中存在的不确定性、难以自适应地确定隐含层神经元,本文提出了一个基于高维云模型和改进RBF神经网络的不确定性分类技术。利用高维正态云创建隐含层神经元,使RBF神经网络能充分表达影像分类过程中存在的不确定性。通过峰值法云变换和高维云算法自适应地确定最优隐含层神经元。通过基于概率的权值确定和频率阈值调整,进一步优化RBF神经网络的结构。实验表明,本文提出的方法有较高的分类精度,分类结果基本上与人眼目视解译一致。 相似文献
4.
Many places around the world are exposed to tropical cyclones and associated storm surges. In spite of massive efforts, a great number of people die each year as a result of cyclone events. To mitigate this damage, improved forecasting techniques must be developed. The technique presented here uses artificial neural networks to interpret NOAA-AVHRR satellite images. A multi-layer neural network, resembling the human visual system, was trained to forecast the movement of cyclones based on satellite images. The trained network produced correct directional forecast for 98% of test images, thus showing a good generalization capability. The results indicate that multi-layer neural networks could be further developed into an effective tool for cyclone track forecasting using various types of remote sensing data. Future work includes extension of the present network to handle a wide range of cyclones and to take into account supplementary information, such as wind speeds, water temperature, humidity, and air pressure. 相似文献
5.
In the proposed approach, the well-known enhanced Lee filter is modified to allow the integration of feature outlines-previously extracted from segmented optical images. The filter is applied to several ENVISAT ASAR images that cover urban, agricultural, and forest areas during different plant phenological stages. The performance of this segment-based speckle filter is compared to those of other filters using ratio images, visual interpretation, and statistical indexes. The approach reduces the loss of radiometry and spatial information. It performs comparable to more complex methods and outperforms common techniques 相似文献
6.
影像匹配在多种计算机视觉任务中起着重要的作用.提出一种用于影像匹配的多尺度视觉相似度比较网络.该网络基于孪生网络结构进行构建,将普通卷积与空洞卷积进行融合,使得卷积神经网络在多个尺度上提取到的视觉特征实现互补.首先,网络的输入由两幅待匹配的影像组成,使用权值共享的两个网络分支分别提取两幅影像的深度特征;其次,在网络中与... 相似文献
7.
联合卷积神经网络与集成学习的遥感影像场景分类 总被引：1，自引：0，他引：1
针对人工设计的中、低层特征难以实现复杂场景影像的高精度分类以及卷积神经网络依赖大量训练数据等问题，结合迁移学习与集成学习，提出了一种联合卷积神经网络与集成学习的遥感影像场景分类算法。首先基于迁移学习的思想，利用在自然影像数据集上训练好的多个深层卷积神经网络模型作为特征提取器，提取图像多个高度抽象的语义特征；然后构建由Logistic回归和支持向量机组成的Stacking集成模型，对同一图像的多个特征分别训练Logistic模型，将预测概率结果融合构建概率特征；最后利用支持向量机对概率特征训练和预测，得到场景影像的分类结果。利用UCMerced_LandUse和NWPU-RESISC 45两种不同规模的遥感影像数据集进行试验，即使在只有10%的数据作为训练样本情况下，本文方法能够分别达到90.74%和87.21%的分类精度。 相似文献
8.
通过对GF-2卫星影像正射校正及波段模拟配准误差试验,分析GF-2卫星正射校正方法的选择以及不同配准误差下对GF-2卫星影像自动分类结果的影响;最后介绍GF-2遥感影像在森林资源监测应用中的初步测试。研究结果表明:正射校正时,当校正精度要求控制在RMS2时,控制点数量选择范围在85~95间较为合理,且控制点数在90个时,RMS值最小;经有理函数模型与卫片模型比较后,卫片模型校正精度较高;以目视判读为主时,实践中建议使用三次卷积重采样法输出结果最好;波段模拟配准误差试验中,配准误差与各地类面积变化间存在显著的线性关系;对于森林面积监测时,配准误差应小于0.3个像元。此研究可为新型国产卫星数据在森林资源监测中的应用提供参考。 相似文献
9.
为了实现大范围湿地动态监测,本文以辽河入海口附近的盘锦湿地为研究区,基于Sentinel2-L1C和AW3D30 DSM数据,在随机森林分类的基础上,结合地形数字特征和多边形形状特征对研究区进行湿地信息提取。通过人工目视解译对该分类方法进行精度验证,结果表明：该方法的自动化程度较高,能够在较少的人工干预下提取湿地覆盖范围。提取结果精度较高,制图精度和总体精度分别为91.04%和82.65%,Kappa系数为0.599 7,说明本文所采用的计算机分类方法与人工目视解译方法具有较好的一致性。 相似文献
10.
针对无人机影像深度学习分类方法缺乏现状,本文利用深度学习理论卷积神经网络方法对无人机影像进行了分类。该法首先抽取无人机影像作为训练集和检验集,然后建立一个2个卷积层-池化层的卷积神经网络模型进行深度学习,通过设定参数并运行模型实现无人机影像分类。实验表明,本文提出的方法可完成较复杂地区无人机影像分类,其分类精度与支持向量机方法相当,为无人机遥感影像分类提供了一个崭新的技术视点。 相似文献
11.
高分辨率遥感影像包含丰富的土地利用类型信息,针对单一卷积神经网络提取图像特征信息不足的问题,提出了一种多结构卷积神经网络（convolutional neural network,CNN）特征级联的分类方法。首先,选择CaffeNet（convolutional architecture for fast feature embedding）、VGG-S（visual geometry group-slow）、VGG-F（visual geometry group-fast）为实验初始模型,对网络全连接层进行参数微调,采用随机梯度下降法（stochasticgradient descent,SGD）更新网络的权重;然后以微调后的网络分别作为特征提取器对图像提取特征,级联上述3种网络的第二个全连接层输出特征作为图像表达;最后,以多类最优边界分配机（multi-class optimal margindistribution machine,mcODM）获得最终分类结果。实验采用UC Merced land-use数据集进行分类效果检验,结果表明,多结构卷积神经网络级联的方法能够达到97.55%的总体分类精度,相较于CaffeNet、VGG-S和VGG-F等,分类精度分别提升了5.71%、2.72%和5.1%。因此多结构卷积神经网络特征级联的方法能够有效提取目标特征信息,提升土地利用分类精度。 相似文献
12.
Approaches for delineating landslide hazard areas using different training sites in an advanced artificial neural network model 总被引：10，自引：0，他引：10
The current paper presents landslide hazard analysis around the Cameron area, Malaysia, using advanced artificial neural networks
with the help of Geographic Information System (GIS) and remote sensing techniques. Landslide locations were determined in
the study area by interpretation of aerial photographs and from field investigations. Topographical and geological data as
well as satellite images were collected, processed, and constructed into a spatial database using GIS and image processing.
Ten factors were selected for landslide hazard including: 1) factors related to topography as slope, aspect, and curvature;
2) factors related to geology as lithology and distance from lineament; 3) factors related to drainage as distance from drainage;
and 4) factors extracted from TM satellite images as land cover and the vegetation index value. An advanced artificial neural
network model has been used to analyze these factors in order to establish the landslide hazard map. The back-propagation
training method has been used for the selection of the five different random training sites in order to calculate the factor’s
weight and then the landslide hazard indices were computed for each of the five hazard maps. Finally, the landslide hazard
maps (five cases) were prepared using GIS tools. Results of the landslides hazard maps have been verified using landslide
test locations that were not used during the training phase of the neural network. Our findings of verification results show
an accuracy of 69%, 75%, 70%, 83% and 86% for training sites 1, 2, 3, 4 and 5 respectively. GIS data was used to efficiently
analyze the large volume of data, and the artificial neural network proved to be an effective tool for landslide hazard analysis.
The verification results showed sufficient agreement between the presumptive hazard map and the existing data on landslide
areas. 相似文献
13.
ABSTRACTTree species distribution mapping using remotely sensed data has long been an important research area. However, previous studies have rarely established a comprehensive and efficient classification procedure to obtain an accurate result. This study proposes a hierarchical classification procedure with optimized node variables and thresholds to classify tree species based on high spatial resolution satellite imagery. A classification tree structure consisting of parent and leaf nodes was designed based on user experience and visual interpretation. Spectral, textural, and topographic variables were extracted based on pre-segmented images. The random forest algorithm was used to select variables by ranking the impact of all variables. An iterating approach was used to optimize variables and thresholds in each loop by comprehensively considering the test accuracy and selected variables. The threshold range for each selected variable was determined by a statistical method considering the mean and standard deviation for two subnode types at each parent node. Classification of tree species was implemented using the optimized variables and thresholds. The results show that (1) the proposed procedure can accurately map the tree species distribution, with an overall accuracy of over 86% for both training and test stages; (2) critical variables for each class can be identified using this proposed procedure, and optimal variables of most tree plantation nodes are spectra related; (3) the overall forest classification accuracy using the proposed method is more accurate than that using the random forest (RF) and classification and regression tree (CART). The proposed approach provides results with 3.21% and 7.56% higher overall land cover classification accuracy and 4.68% and 10.28% higher overall forest classification accuracy than RF and CART, respectively. 相似文献
14.
针对城市行道树调查中,街景影像背景环境复杂多变、行道树个体差异大,依靠目视判读费时费力的问题,该文基于车载移动测量系统采集的全景影像数据,利用深度学习算法,在快速区域卷积神经网络的目标检测方法基础上,建立适用于街景行道树检测的深度神经网络模型。模型采用基于共有显著性区域及冗余策略的行道树多示例目标候选区域选择方法,使用车载图像的几何约束进一步筛选合适的候选区域,从而实现行道树目标候选区域的统一选择,提升行道树目标的检测效果。实验结果表明,该文提出的方法能够实现多种行道树的准确自动识别与提取,进而大大降低行道树绿化调查的成本。 相似文献
15.
联合显著性和多层卷积神经网络的高分影像场景分类 总被引：3，自引：2，他引：1
高分辨率遥感影像中的场景信息,对于影像解译和现实世界的理解具有重要意义。传统的场景分类方法多利用中、低层人工特征,但是高分辨率遥感影像的信息丰富,场景构成复杂,需要高层次的特征来表达。本文提出了一种联合显著性和多层卷积神经网络的方法,首先利用显著性采样获取包含影像主要信息的有意义的块,将这些块作为样本集输入卷积神经网络中进行训练,获得不同层次的特征表达,最后联合多层特征利用支持向量机进行分类。两组高分影像场景数据UC Merced 21类和Wuhan 7类试验表明,显著性采样能够有效地获取主要目标,减弱其他无关目标的影响,降低数据冗余;卷积神经网络能够自动学习高层次的特征,相比已有方法,本文方法能够有效提高分类精度。 相似文献
16.
Forest canopy density stratification using biophysical modeling 总被引：1，自引：0，他引：1
Forest canopy density is an important parameter to assess the ecological conditionsviz, light penetration through canopy, undergrowth, surface reflectance, rainfall interception, etc. in a forest landscape. The rate of change in the cover and density has increased due to human need for fuel and fodder. Hence, quick, repetitive and accurate information about forest density is required at the local, regional, state and national levels for sustainable forest management. Satellite remote sensing has the potential to provide information on the forest canopy closure. The present study aims at forest canopy density mapping using satellite remote sensing data using three techniques: visual interpretation (VI), object oriented image segmentation (OOIS) and biophysical modeling (BM). On comparing the techniques, the BM has been found to be the better density mapping technique than other two in terms of accuracy, efficiency and high correlation with ground estimates. 相似文献
17.
NOAA/AVHRR Global Vegetation Index (GVI) data of Asia in 1983 and 1987 were used to evaluate their usefulness for global land cover monitoring. Color composite images of monthly GVI data and color composite images of principal components from 12 successive monthly GVI data were found to be useful for visual interpretation of seasonal vegetation dynamics. The results of cluster analysis applied to monthly GVI data for a one‐year period, indicate that unsupervised classification method is useful for global or continental land cover classification without ground truth. In order to detect land cover changes, the difference between the 1983 and 1987 12‐month GVI data was calculated. The results show that it is difficult to detect land cover changes due to cloud contamination in monthly GVI data and poor registration of GVI products. 相似文献
18.
The current paper presents landslide hazard analysis around the Cameron area, Malaysia, using advanced artificial neural networks with the help of Geographic Information System (GIS) and remote sensing techniques. Landslide locations were determined in the study area by interpretation of aerial photographs and from field investigations. Topographical and geological data as well as satellite images were collected, processed, and constructed into a spatial database using GIS and image processing. Ten factors were selected for landslide hazard including: 1) factors related to topography as slope, aspect, and curvature; 2) factors related to geology as lithology and distance from lineament; 3) factors related to drainage as distance from drainage; and 4) factors extracted from TM satellite images as land cover and the vegetation index value. An advanced artificial neural network model has been used to analyze these factors in order to establish the landslide hazard map. The back-propagation training method has been used for the selection of the five different random training sites in order to calculate the factor’s weight and then the landslide hazard indices were computed for each of the five hazard maps. Finally, the landslide hazard maps (five cases) were prepared using GIS tools. Results of the landslides hazard maps have been verified using landslide test locations that were not used during the training phase of the neural network. Our findings of verification results show an accuracy of 69%, 75%, 70%, 83% and 86% for training sites 1, 2, 3, 4 and 5 respectively. GIS data was used to efficiently analyze the large volume of data, and the artificial neural network proved to be an effective tool for landslide hazard analysis. The verification results showed sufficient agreement between the presumptive hazard map and the existing data on landslide areas. 相似文献
19.
Anto A. Micheal K. Vani S. Sanjeevi 《Journal of the Indian Society of Remote Sensing》2014,42(2):445-451
Lunar surface exploration is increasing rapidly with priority being given to precision of lunar soft landing. Lunar soft landing is achieved when craters and grabens are used as navigational landmarks. Grabens are formed from localized tensional stress fields or from near-surface dike emplacement. These tectonic features tend to have consistently straight or accurate parallel-striking walls bounded by steep, inward-dipping normal faults. Aiming at navigational application, a novel approach for automatic graben detection based on Hessian technique has been implemented on Digital Terrain Model (DTM) of lunar image. The Hessian technique uses gradient change as a key parameter to identify grabens. Adaptive Binarization using Otsu method is used to extract graben features from the Hessian image. Features such as small grabens and craters are removed using morphological operations, resulting in significant appearance of grabens. The experiment is conducted in different DTM images of lunar surface and the results indicate 90 % of the grabens are detected. The statistical results are evaluated based on visual interpretation, for both automatic and manual graben detection. It is observed that the proposed automatic graben detection technique gives better results than the manual detection. 相似文献
20.
The present study has generated spatial databases on the vegetation type with plant biodiversity, forest fragmentation and
disturbance regimes in Tamilnadu parts of Eastern Ghats (EG), India. These databases have been analysed geospatially with
landscape ecology approach. The study also includes ground inventory of plant species based on Remote Sensing (RS) data stratification.
The vegetation type map was generated from the visual interpretation of two season IRS LISS III data. The spatial landscape
analysis of the remotely sensed interpreted images was carried out using customized software, SPLAM. This is first such study
in Tamilnadu Eastern Ghats that provides a comprehensive spatial database on vegetation types, disturbance regime and plant
species diversity. The study has shown that the dry deciduous and thorn forests have shown better resistance to disturbance
compared to the most disturbed evergreen and semievergreen forests. The study outputs are being utilized by forest department
and biodiversity boards for conservation action planning and compliance to Convention on Biological Diversity (CBD). 相似文献