排序方式: 共有133条查询结果,搜索用时 203 毫秒
51.
近年来,新兴机载全波形Li DAR技术在摄影测量与遥感领域受到广泛关注。其相比传统机载Li DAR的优势在于它对后向散射回波信息进行了完整的数字化记录。波形分解是波形数据处理的重要环节,通过波形分解不仅仅可以得到高质量的三维点云产品,还能够提高DEM的生成精度并辅助点云的精确分类,随着研究的不断深入,全波形数据将在更多领域发挥更大的实用效益。 相似文献
52.
53.
Over the past 20 years, sonar imaging technology particularly for the high-technology sector has been a focus of research, in which many developed countries, especially those with coast lines, have been competing with each other. It has seen a rapid development with increasing widespread applications that has played an important and irreplaceable role in underwater exploration with great prospects for social, economic, scientific, and military benefits. The fundamental techniques underlying sonar imaging, including multi-beamforming, synthetic-aperture and inverse synthetic-aperture sonar, acoustic lensing, and acoustical holography, are described in this paper. This is followed by a comprehensive and systematic review on the advantages and disadvantages of these imaging techniques, applicability conditions, development trends, new ideas, new methods, and improvements in old methods over recent years with an emphasis on the situation in China, along with a bold and constructive prediction to some development characteristics of sonar imaging technology in the near future in China. The perspectives presented in this paper are offered with the idea of providing some degree of guidance and promotion of research on sonar imaging technology. 相似文献
54.
55.
56.
级联卷积神经网络的遥感影像飞机目标检测 总被引:1,自引:0,他引:1
传统遥感影像飞机目标检测算法依赖于人工设计特征,对大范围复杂场景和多尺度的飞机目标稳健性较差,基于深层卷积神经网络的目标检测算法通常难以有效应对大幅影像的目标搜索和弱小目标检测问题,针对上述问题,本文提出了一种基于级联卷积神经网络的遥感影像飞机目标检测算法。首先根据全卷积神经网络能够支持输入任意大小图像的特点,采用小尺度浅层全卷积神经网络对整幅影像进行遍历和搜索,快速获取疑似飞机目标作为兴趣区域,然后利用较深层的卷积神经网络对兴趣区域进行更精确的目标分类与定位。为提高卷积神经网络对地物目标的辨识能力,在卷积层中引入多层感知器,并在训练过程中采取多任务学习与离线难分样本挖掘的策略;在测试阶段,建立影像金字塔进行多级搜索,并结合非极大值抑制消除冗余窗口,从而实现由粗到精的飞机目标检测与识别。对多个数据集下多种复杂场景的遥感影像进行测试,结果表明,本文方法具有较高的准确性和较强的稳健性,可为大幅遥感影像的飞机目标检测问题提供一个快速高效的解决方案。 相似文献
57.
58.
59.
60.