当前位置: X-MOL 学术Geocarto Int. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Accurate Building Imaging from VHR Imagery Using Generalized Differential Morphological Attribute Profile
Geocarto International ( IF 3.3 ) Pub Date : 2021-04-23 , DOI: 10.1080/10106049.2021.1920631
Jiannong Cao 1, 2 , Junjun Li 3 , Nan Chen 2 , Muleta Ebissa Feyissa 2
Affiliation  

ABSTRACT

Differential morphological profile (DMP) and differential attribute profile (DAP) are effective tools for automated building extraction from very high-resolution (VHR) images. Compared to the DMP, DAP provides a more flexible way to model image features not only related to the scale of regions, but also any measure (structural, spatial, and spectral) that can be evaluated in the region. The DAP series can be regarded as a structural spectrum that depicts the responses of the image components related to different attributes and scales. However, the conventional DAP-based building feature extraction approach ignores discriminative information for features that are across the scales in the attribute profiles. In this study, an improvement in the discriminatory power of DAP for building feature extraction is proposed. To obtain the entire differential profiles, a scale-span difference attribute profile called generalized DAP (GDAP) is presented, based on which, a new generalized morphological attribute-building index is proposed for the automatic extraction of buildings from VHR imagery. The GDAP can describe the complete attribute spectrum and measure the difference between arbitrary scales, which is more appropriate for representing the multiscale characteristics of buildings. The performance of the proposed method was validated using VHR datasets from an aerial platform. In terms of statistical precision and visual inspection, the results of the proposed framework are superior to those of the relevant supervised and unsupervised building extraction approaches.



中文翻译:

使用广义差分形态属性配置文件从VHR图像进行准确的建筑物成像

摘要

差异形态特征(DMP)和差异属性特征(DAP)是从超高分辨率(VHR)图像中自动提取建筑物的有效工具。与DMP相比,DAP提供了一种更为灵活的方法来对图像特征进行建模,该特征不仅与区域的比例有关,而且还可以对区域中可以评估的任何度量(结构,空间和光谱)进行建模。DAP系列可视为描述与不同属性和比例有关的图像成分响应的结构光谱。但是,传统的基于DAP的建筑特征提取方法会忽略针对属性概要中各个尺度的特征的判别信息。在这项研究中,提出了DAP对建筑物特征提取的鉴别能力的改进。为了获得整个差分轮廓,提出了一种称为通用DAP(GDAP)的比例尺差异属性轮廓,在此基础上,提出了一种新的广义形态属性构建索引,用于从VHR图像中自动提取建筑物。GDAP可以描述完整的属性谱并测量任意尺度之间的差异,这更适合于表示建筑物的多尺度特征。使用空中平台的VHR数据集验证了该方法的性能。在统计精度和视觉检查方面,所提出框架的结果优于相关的有监督和无监督建筑物提取方法。在此基础上,提出了一种新的广义形态属性建立指标,用于从VHR图像中自动提取建筑物。GDAP可以描述完整的属性谱并测量任意尺度之间的差异,这更适合于表示建筑物的多尺度特征。使用空中平台的VHR数据集验证了该方法的性能。在统计精度和视觉检查方面,所提出框架的结果优于相关的有监督和无监督建筑物提取方法。在此基础上,提出了一种新的广义形态属性建立指标,用于从VHR图像中自动提取建筑物。GDAP可以描述完整的属性谱并测量任意尺度之间的差异,这更适合于表示建筑物的多尺度特征。使用空中平台的VHR数据集验证了该方法的性能。在统计精度和视觉检查方面,所提出框架的结果优于相关的有监督和无监督建筑物提取方法。GDAP可以描述完整的属性谱并测量任意尺度之间的差异,这更适合于表示建筑物的多尺度特征。使用空中平台的VHR数据集验证了该方法的性能。在统计精度和视觉检查方面,所提出框架的结果优于相关的有监督和无监督建筑物提取方法。GDAP可以描述完整的属性谱并测量任意尺度之间的差异,这更适合于表示建筑物的多尺度特征。使用空中平台的VHR数据集验证了该方法的性能。在统计精度和视觉检查方面,所提出框架的结果优于相关的有监督和无监督建筑物提取方法。

更新日期:2021-04-23
down
wechat
bug