【摘要】：衛(wèi)星影像的RFM模型具有傳感器無關(guān)的優(yōu)點(diǎn),適用于多源影像的幾何定位處理,但在無地面控制點(diǎn)條件下聯(lián)合定位時(shí)存在自主定位優(yōu)勢(shì)影像難以發(fā)揮主導(dǎo)作用且求解易發(fā)散的不足。本文通過將影像的先驗(yàn)自主定位精度和成像線性漂移轉(zhuǎn)化為像方定向參數(shù)的精度和權(quán)信息,建立考慮影像定向參數(shù)精度信息的RFM模型。以12景TerraSAR-X和6對(duì)12景覆蓋面積約為18萬(wàn)km2的SPOT-5HRS立體長(zhǎng)條帶影像為數(shù)據(jù)源,對(duì)兩類影像定向參數(shù)先驗(yàn)精度配置偏差、SAR影像升降軌道方向、SAR影像數(shù)目、SAR影像分布等因素對(duì)定位精度的影響進(jìn)行了系列定位試驗(yàn),少量SAR與大范圍HRS聯(lián)合的影像自主定位平面/高程精度可達(dá)6.0m/4.2m。本文RFM平差模型無地面控制點(diǎn)定位精度和定向參數(shù)求解穩(wěn)健性相對(duì)于傳統(tǒng)模型有顯著提升,是衛(wèi)星影像無控制點(diǎn)1∶10萬(wàn)/1∶5萬(wàn)全球測(cè)圖的一種潛在方法。
[Abstract]:The RFM model of satellite images has the advantages of sensor-independent, which is suitable for geometric location processing of multi-source images. However, in the absence of ground control points (GCPs), the dominant image of autonomous positioning is difficult to play a dominant role in joint positioning, and the solution is easy to diverge. By transforming the prior autonomous positioning accuracy and the linear drift of the image into the accuracy and weight information of the orientation parameters of the image, a RFM model considering the precision information of the orientation parameters of the image is established in this paper. Taking 12-scene TerraSAR-X and 6 pairs of SPOT-5HRS long strip images covering about one hundred and eighty thousand km2 as the data source, the prior accuracy deviation of the two types of image orientation parameters, the rise-and-fall orbit direction of SAR images and the number of SAR images, are used as data sources for the two types of image orientation parameters. The influence of SAR image distribution and other factors on the positioning accuracy has been carried out in a series of positioning experiments. The autonomous positioning plane / elevation accuracy of a small amount of SAR combined with a large range of HRS images can reach 6.0 m / m 路m ~ (- 1). In this paper, the accuracy of positioning without ground control points and the robustness of orientation parameters of the RFM adjustment model are significantly improved compared with the traditional model. It is a potential method for global mapping of satellite images without control points 1: one hundred thousand / 1: 50 000.
【作者單位】： 中國(guó)測(cè)繪科學(xué)研究院;
【基金】：國(guó)土資源部公益行業(yè)科研專項(xiàng)(201411119-2)~~
【分類號(hào)】：P237
，

本文編號(hào)：2459792