摘要
随着现代科学技术的发展,大数据已经成为这个时代的标志。因此快速、准确地采集图像数据变得十分重要。本文研究的是实时线段检测算法EDLines,首先通过实验对比分析及所给公式,可以获得梯度幅度和方向、线验证参数、梯度阈值、锚点阈值和扫描间隔参数值,基于所获得的参数值,在处理不同类型图像时不需要改变其参数就能获取准确的结果,然后对所研究的算法EDLines与现今已知运行较快的直线提取算法进行对比分析,得到本文所研究的EDLines算法相比较其他直线提取算法都具有最优性。最后,根据本文研究的EDLines算法对遥感影像进行直线提取,获得清晰且假阳性较少的直线提取图像。超高的精度和计算速度使实时计算机视觉和图像处理的应用迈出了崭新的一步。
With the development of modern science and technology, large data has become a symbol of this era. Therefore, fast and accurate acquisition of image data becomes very important. In this paper, a real-time line detection algorithm, called EDLines, is proposed. Firstly, through comparative analysis of the experiment and the given formula, the gradient magnitude and direction, line verification parameters, gradient threshold, anchor threshold and scanning interval parameter values can be obtained. Based on the obtained parameter values, accurate results can be obtained without changing their parameters when dealing with different types of images. Then compare the algorithm EDLines with the current fast line extraction algorithm, and get the EDLines algorithm studied in this paper. Compared with other linear extraction algorithms, EDLines have the optimality. Finally, according to the algorithm, the remote sensing image is extracted by straight line, and the image is extracted with clear and false positive. With ultra-high accuracy and high speed, it will be an new step for next-generation real-time computer vision and image processing applications.
作者
齐永菊
裴亮
魏显虎
张宗科
李航
QI Yongju;PEI Liang;WEI Xianhu;ZHANG Zongke;LI Hang(Institute of Remote Sensing and Digital Earth,Chinese Academy of Science,Beijing 100101,China;School of Geomatics,Liaoning Technical University,Fuxin 123000,China;Sino-African Joint Research Center,CAS,Wuhan 430000,China)
出处
《测绘与空间地理信息》
2018年第9期109-113,共5页
Geomatics & Spatial Information Technology
基金
国家自然科学基金(41401068)
中国科学院境外机构建设项目(SAJC201608)
东非生态环境及生物多样性保护空间大数据共享与服务平台(SAJC201608)资助
