• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Journal of geovisualization and spatial analysis >Hybrid Behrens-Fisher- and Gray Contrast-Based Feature Point Selection for Building Detection from Satellite Images
【24h】

Hybrid Behrens-Fisher- and Gray Contrast-Based Feature Point Selection for Building Detection from Satellite Images

机译：Hybrid Behrens-Fisher- and Gray Contrast-Based Feature Point Selection for Building Detection from Satellite Images

获取原文
获取原文并翻译 | 示例
       
页面导航
  • 摘要
  • 著录项
  • 相关主题

摘要

Over the past few years, satellite images have been one of the most influential and paramount tools utilized by meteorologists since these images soothe forecasters with a comprehensible, crisp, and correct representation of evolving events. Moreover, the satellite images acquired from remote sensing are a quicker method to detect any feature both locally and globally. Building detection remains to be one of the most challenging issues as buildings have to be differentiated between non-building objects. In this paper, we designed a hybrid feature selection process using the hybrid Behrens-Fisher- and gray contrast-based feature point selection method. The hybrid Behrens-Fisher- and gray contrast-based feature point selection (HBF-GCFPS) increases the extraction accuracy building detection efficiently by detecting feature key points and match points based on oriented features in the oriented features from accelerated segment test and rotated binary robust independent elementary features (ORB) algorithm. We first applied the Max mean filter-based preprocessing to the raw Massachusetts building dataset based on tone and association, which not only removes noise but also preserves geometrical features. Next, we designed the Behrens-Fisher-based feature point description using binary robust independent elementary features (BRIEF) with the aid of size and shape, which obtains the feature points robustly and accurately. Finally, we designed the gray contrast-based feature point extraction using features from accelerated segment test (FAST) with the aid of an arbitrary threshold calculation model, improvising the method's potentiality to extract feature points at analogous regions, where some small building structures are situated closely with minimum time. Experiment findings indicate that the HBF-GCFPS techniques can achieve better performance than other latest methods in terms of training time, false-positive rate, time per frame, match rate key points, matching time, and average key points.

著录项

获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号