Coverage of nominal 95% confidence intervals of a proportion estimated from a sample obtained under a complex survey design, or a proportion estimated from a ratio of two random variables, can depart significantly fro...Coverage of nominal 95% confidence intervals of a proportion estimated from a sample obtained under a complex survey design, or a proportion estimated from a ratio of two random variables, can depart significantly from its target. Effective calibration methods exist for intervals for a proportion derived from a single binary study variable, but not for estimates of thematic classification accuracy. To promote a calibration of confidence intervals within the context of land-cover mapping, this study first illustrates a common problem of under and over-coverage with standard confidence intervals, and then proposes a simple and fast calibration that more often than not will improve coverage. The demonstration is with simulated sampling from a classified map with four classes, and a reference class known for every unit in a population of 160,000 units arranged in a square array. The simulations include four common probability sampling designs for accuracy assessment, and three sample sizes. Statistically significant over- and under-coverage was present in estimates of user’s (UA) and producer’s accuracy (PA) as well as in estimates of class area proportion. A calibration with Bayes intervals for UA and PA was most efficient with smaller sample sizes and two cluster sampling designs.展开更多
针对现有基于深度学习的遥感影像分割方法难以充分考虑像素之间关系,而全连接条件随机场(fully connected conditional random fields,FullCRF)后处理效率低下且难以训练的问题,提出了结合改进金字塔场景解析网络(pyramid scene parsing...针对现有基于深度学习的遥感影像分割方法难以充分考虑像素之间关系,而全连接条件随机场(fully connected conditional random fields,FullCRF)后处理效率低下且难以训练的问题,提出了结合改进金字塔场景解析网络(pyramid scene parsing network,PSPNet)算法与卷积条件随机场(convolutional condition random fields,ConvCRF)的方法。首先,在PSPNet中采用更加密集连接的DenseNet网络,并在高低层特征融合部分将原有的连接CNN网络末端特征图方式改为连接第三个dense模块。其次,在改进PSPNet基础上,设计与ConvCRF的集成方法,通过引入两个损失函数,设计两步法训练方式,实现了集成模型的端对端训练。最后,进行某区域无人机遥感影像4类要素分割及马萨诸塞州航空遥感影像道路分割试验。结果表明,改进PSPNet在无人机影像分割试验中MIoU(mean intersection over union)提升0.25%,总体精度提升0.47%;结合ConvCRF处理模块后,MIoU可进一步提升0.94%,总体精度进一步提升0.47%,单幅图像计算时间仅增加79 ms,且精度优于FullCRF,时间开销仅为FullCRF的35%,在马萨诸塞州道路分割试验中,本方法较其他精度更优。展开更多
This study has tried to prove the ability of remote sensing techniques to extract information necessary for preparation of geological mapping of the earth’s surface using multi-spectral satellite images which are ric...This study has tried to prove the ability of remote sensing techniques to extract information necessary for preparation of geological mapping of the earth’s surface using multi-spectral satellite images which are rich sources of Earth’s surface information. In this study, the surface geological mappings of Zefreh region have been investigated through ASTER, OLI, and IRS-PAN remote sensing data. To prepare the geological map, preprocessing steps and reducing noises from data using MNF algorithm were firstly carried out. Then a set of processing algorithms and image classification methods are included;the band rationing, color composite and pixel classification based on maximum likelihood, spectral and sub-pixel classification methods of spectral angle mapper (SAM), spectral feature fitting (SFF), linear spectral differentiation (LSU), hill-shade images and automatic lineament extraction were used. Confusion matrix was formed for all classified images through control points were randomly selected from 1:25,000 map of the region to determine the accuracy of obtained results, which indicated the maximum accuracy (up to 90%) of output images. Comparing the results obtained from these methods with the map prepared by ground operations confirmed accuracy results. Finally, the surface geology and fault map of Zafreh region was produced by combining detected geological formations and tectonic lineaments.展开更多
Remote sensing images show a very promising perspective for distinguishing tree species,especially those with the very high resolution ranging from 1 to 4 m.However,the traditional methodology for classifying land cov...Remote sensing images show a very promising perspective for distinguishing tree species,especially those with the very high resolution ranging from 1 to 4 m.However,the traditional methodology for classifying land cover types,solely depending on spectral features,while texture and other spatial information are neglected, has the weakness such as inadequately utilization of information,low accuracies of classification,etc. Considering to the texture differences among forest species,it is more important for spatial information description of high-resolution remote sensing image to improve the precision of textural features choosing.In this study,the factors to influence the nine textural features choosing were analyzed and the results showed that the moving window size was the main factor to affect the obtaining processes of textural features based on the gray level co-occurrence matrix(GLCM) method,and the imagery was then classified combining the maximum likelihood classification(MLC) method with the original spectral values and texture features.First,this study utilized a correlation analysis of the images from a principal component analysis.Second,through multiple information sources,including textual features derived from the data.For the high-resolution remote sensing image, the most proper moving window size was determined from 3×3 to 31×31.Classification of the major tree species throughout the study area (the SunYat-Sen Mausoleum in Nanjing) was undertaken using the MLC.Third,to aid forest research,classification accuracy was improved using the GLCM.According to correlations among textures and richness of the data,GLCM provided the best window size and textural parameters. Results indicated that the texture characteristics were add in the spectral characteristics to improve the precision of the results of the classification, 19×19 window for best window.The total precision can reach 66.322 6%,Kappa coefficient is 0.584 0.Each tree species has greatly improved accuracies of the classification.By the calc展开更多
文摘Coverage of nominal 95% confidence intervals of a proportion estimated from a sample obtained under a complex survey design, or a proportion estimated from a ratio of two random variables, can depart significantly from its target. Effective calibration methods exist for intervals for a proportion derived from a single binary study variable, but not for estimates of thematic classification accuracy. To promote a calibration of confidence intervals within the context of land-cover mapping, this study first illustrates a common problem of under and over-coverage with standard confidence intervals, and then proposes a simple and fast calibration that more often than not will improve coverage. The demonstration is with simulated sampling from a classified map with four classes, and a reference class known for every unit in a population of 160,000 units arranged in a square array. The simulations include four common probability sampling designs for accuracy assessment, and three sample sizes. Statistically significant over- and under-coverage was present in estimates of user’s (UA) and producer’s accuracy (PA) as well as in estimates of class area proportion. A calibration with Bayes intervals for UA and PA was most efficient with smaller sample sizes and two cluster sampling designs.
文摘针对现有基于深度学习的遥感影像分割方法难以充分考虑像素之间关系,而全连接条件随机场(fully connected conditional random fields,FullCRF)后处理效率低下且难以训练的问题,提出了结合改进金字塔场景解析网络(pyramid scene parsing network,PSPNet)算法与卷积条件随机场(convolutional condition random fields,ConvCRF)的方法。首先,在PSPNet中采用更加密集连接的DenseNet网络,并在高低层特征融合部分将原有的连接CNN网络末端特征图方式改为连接第三个dense模块。其次,在改进PSPNet基础上,设计与ConvCRF的集成方法,通过引入两个损失函数,设计两步法训练方式,实现了集成模型的端对端训练。最后,进行某区域无人机遥感影像4类要素分割及马萨诸塞州航空遥感影像道路分割试验。结果表明,改进PSPNet在无人机影像分割试验中MIoU(mean intersection over union)提升0.25%,总体精度提升0.47%;结合ConvCRF处理模块后,MIoU可进一步提升0.94%,总体精度进一步提升0.47%,单幅图像计算时间仅增加79 ms,且精度优于FullCRF,时间开销仅为FullCRF的35%,在马萨诸塞州道路分割试验中,本方法较其他精度更优。
文摘This study has tried to prove the ability of remote sensing techniques to extract information necessary for preparation of geological mapping of the earth’s surface using multi-spectral satellite images which are rich sources of Earth’s surface information. In this study, the surface geological mappings of Zefreh region have been investigated through ASTER, OLI, and IRS-PAN remote sensing data. To prepare the geological map, preprocessing steps and reducing noises from data using MNF algorithm were firstly carried out. Then a set of processing algorithms and image classification methods are included;the band rationing, color composite and pixel classification based on maximum likelihood, spectral and sub-pixel classification methods of spectral angle mapper (SAM), spectral feature fitting (SFF), linear spectral differentiation (LSU), hill-shade images and automatic lineament extraction were used. Confusion matrix was formed for all classified images through control points were randomly selected from 1:25,000 map of the region to determine the accuracy of obtained results, which indicated the maximum accuracy (up to 90%) of output images. Comparing the results obtained from these methods with the map prepared by ground operations confirmed accuracy results. Finally, the surface geology and fault map of Zafreh region was produced by combining detected geological formations and tectonic lineaments.
文摘Remote sensing images show a very promising perspective for distinguishing tree species,especially those with the very high resolution ranging from 1 to 4 m.However,the traditional methodology for classifying land cover types,solely depending on spectral features,while texture and other spatial information are neglected, has the weakness such as inadequately utilization of information,low accuracies of classification,etc. Considering to the texture differences among forest species,it is more important for spatial information description of high-resolution remote sensing image to improve the precision of textural features choosing.In this study,the factors to influence the nine textural features choosing were analyzed and the results showed that the moving window size was the main factor to affect the obtaining processes of textural features based on the gray level co-occurrence matrix(GLCM) method,and the imagery was then classified combining the maximum likelihood classification(MLC) method with the original spectral values and texture features.First,this study utilized a correlation analysis of the images from a principal component analysis.Second,through multiple information sources,including textual features derived from the data.For the high-resolution remote sensing image, the most proper moving window size was determined from 3×3 to 31×31.Classification of the major tree species throughout the study area (the SunYat-Sen Mausoleum in Nanjing) was undertaken using the MLC.Third,to aid forest research,classification accuracy was improved using the GLCM.According to correlations among textures and richness of the data,GLCM provided the best window size and textural parameters. Results indicated that the texture characteristics were add in the spectral characteristics to improve the precision of the results of the classification, 19×19 window for best window.The total precision can reach 66.322 6%,Kappa coefficient is 0.584 0.Each tree species has greatly improved accuracies of the classification.By the calc
