龙恩, 吕守业, 曲小飞, 孟钢, 赖广陵, 杨宇科. 使用同名弧距的卫星影像油罐高度反演模型[J]. 武汉大学学报 ( 信息科学版), 2024, 49(3): 411-418. DOI: 10.13203/j.whugis20210239
引用本文: 龙恩, 吕守业, 曲小飞, 孟钢, 赖广陵, 杨宇科. 使用同名弧距的卫星影像油罐高度反演模型[J]. 武汉大学学报 ( 信息科学版), 2024, 49(3): 411-418. DOI: 10.13203/j.whugis20210239
LONG En, LÜ Shouye, QU Xiaofei, MENG Gang, LAI Guangling, YANG Yuke. Height Inversion Model of Oil Tank Using Satellite Imagery with Same Name Arc Distance[J]. Geomatics and Information Science of Wuhan University, 2024, 49(3): 411-418. DOI: 10.13203/j.whugis20210239
Citation: LONG En, LÜ Shouye, QU Xiaofei, MENG Gang, LAI Guangling, YANG Yuke. Height Inversion Model of Oil Tank Using Satellite Imagery with Same Name Arc Distance[J]. Geomatics and Information Science of Wuhan University, 2024, 49(3): 411-418. DOI: 10.13203/j.whugis20210239

使用同名弧距的卫星影像油罐高度反演模型

Height Inversion Model of Oil Tank Using Satellite Imagery with Same Name Arc Distance

  • 摘要: 为提高卫星影像反演油罐高度的精度,构建了一种使用同名弧距(same name arc distance, SNAD)的圆柱型油罐高度反演模型。首先,定义单幅卫星影像上SNAD概念,明确油罐影像的主要几何要素及其物理意义;然后,结合圆柱型油罐空间结构的各向同性特征,进行成像条件多策略组合的成像模拟,构建高度反演模型;最后,明确模型中各输入参数,推导SNAD核心参数解算方法。选取四型亚米级卫星影像,对两型油罐进行高度反演实验,与常规方法进行对比分析。所提方法能有效地实现多型卫星图像上同名弧边的检测和高度反演,在反演精度方面,所提模型的均方根误差为0.50~1.00 m,平均达0.78 m,而常规模型的均方根误差为9.28~9.59 m,平均仅为9.44 m;在反演的稳定性方面,所提模型标准差为0.35~0.63 m,平均为0.46 m,而常规模型标准差为6.50~6.69 m,平均为6.60 m。结果表明,所提模型反演精度高、稳定性好、普适性强、便捷高效,为开展卫星影像油罐高度反演及精细化判读提供一种新的模型算法。

     

    Abstract:
    Objectives In order to improve the inversion accuracy of cylindrical oil tank height from satellite imagery, this paper introduces a height inversion model of oil tank using satellite imagery with the same name arc distance (SNAD).
    Methods First, we define the concept of SNAD and clarify several main geometric elements and the physical meanings of oil tank on satellite image. Then, considering the isotropic characteristics of the cylindrical oil tank, we propose a height inversion model based on imaging simulation under multiple strategic conditions. Finally, we clarify each parameter in the inversion model and deduce the calculation method of key parameters. In order to verify the accuracy and applicability of the proposed model, we invert the height of two types of oil tanks in four high-resolution satellite images and compare with the conventional methods.
    Results The proposed model can effectively realize the detection of SNAD and height inversion of cylindrical oil tank in multi-satellite images. The root mean square error of the proposed model ranges from 0.50 m to 1.00 m, with an average value of 0.78 m, while that of the conventional model ranges from 9.28 m to 9.59 m, with an average value of 9.44 m. The standard deviation of the proposed model ranges from 0.35 m to 0.63 m, with an average value of 0.46 m, while that of the conventional model ranges from 6.50 m to 6.69 m, with an average value of 6.60 m.
    Conclusions The proposed model has the advantages of high inversion accuracy, excellent stability, strong universality, convenience and efficiency. It can provide a new solution to carry out oil tank height inversion based on high-resolution satellite image.

     

/

返回文章
返回