资源三号测绘卫星为了获取较大的幅宽和较高的空间分辨率,其三线阵相机和多光谱相机采用了多片CCD拼接成像的方式获取地面影像。如直接提供分片的CCD影像产品,用户难以使用。传感器校正产品解决了CCD影像拼接的问题,拼接后的影像不仅视觉无缝,同时几何无缝。该产品同时解决了多光谱波段间配准的问题,且生产过程中引入的投影差误差和交会误差可以忽略不计。该产品消除了大部分畸变,故拥有极高的RFM替代严密成像几何模型精度。本文利用安平和登封地区的三线阵数据进行验证,试验区三线阵传感器校正产品RFM替代精度优于1%像元,无地面控制点立体定位精度优于15m,带控制点平面误差在3m以内,高程误差在2m之内,且三线阵平差平面定位精度要优于两线阵。
To get larger width and higher resolution, ZY-3 gets the ground information by multi-CCD mosaicing. However, it brings much inconvenience to users when operators provide them with divided CCD images. Sensor corrected (SC) products successfully solve the problem of CCD mosaicing, ensuring both visual seamlessness and geometrical seamlessness, the problem of registration among multispectral bands and guarantee the high precision at the same time. Since SC products eliminate most distortions, it has extremely highRFM replacement accuracy of rigorous sensor model. The triplet stereo images of Anping and Dengfeng areas are used to perform the adjustment experiment, in which the RFM replacement RMS (root mean square) errors of rigorous sensor model being less than 1% pixels. When without GCPs (ground control points), the stereo positioning accuracy of bundle adjustment can reach 15 meters while the plane error is less than 3 meters and altitude error less than 2 meters when bundle adjustment is performed with4 corner GCPs. Moreover, the plane positioning accuracy of triplet stereo images is better than stereo images(forward and backward images).