Sensor Modeling for Aerial Triangulation with Three-Line-Scanner (TLS) Imagery

This article describes the sensor modeling and the photogrammetric triangulation procedure for the TLS (Three-Line-Scanner) system. This system is a new airborne digital sensor, developed by STARLABO Corporation, Tokyo jointly with the Institute of Industrial Science, University of Tokyo. It utilizes the Three-Line-Scanner principle to capture digital image triplets in along-strip mode. The imaging system contains three times three (RGB) one-dimensional CCD arrays, with 10 200 pixels of 7μm each, mounted parallel to each other in the focal plane. They produce seamless high-resolution images (5-10 cm footprint on the ground) with three viewing directions (forward, nadir and backward). In order to get precise attitude data and high quality image data from an aerial platform, a high quality stabilizer stabilizes the camera and outputs attitude data at 500 Hz. A Trimble MS750 serves as Rover GPS and collects L1/L2 kinematic data at 5 Hz and another Trimble MS750 serves as Base GPS on the ground. The position and attitude elements measured by the on-board GPS/INS do not refer to the perspective center of the imaging camera. Additionally, there is a boresight misalignment between the axes of the INS and the camera. These translational and rotational offsets have been taken into account in our sensor model and triangulation procedures. In our experiments, the following 3 trajectory model are evaluated: (a) Direct georeferencing with stochastic exterior orientations (DGR), (b) Piecewise Polynomials with kinematic model up to second order and stochastic first and second order constraints (PPM) and (c) Lagrange Polynomials with variable orientation fixes (LIM). With different numbers and distributions of control points and tie points, 4.9-6.3 cm and 8.6-9.4 cm absolute accuracy in planimetry and height is achieved using the DGR model under the condition that the GPS/camera displacement corrections have been applied. Moreover, with different numbers of spline sections or orientation fixes, 2.6-6.0 cm and 4.9-11.7 cm absolute accuracy in planimetry and height is attained using the PPM and LIM models. These results show that a ground point determination of 0.5-1.2 pixel accuracy in planimetry and 0.7-2.1 pixel accuracy in height has been achieved. The orientation parameter determination using the DGR model has the advantage of stability and needs less control points, but the obtained accuracy is better with the PPM and LIM models. This however is penalized by the need to have more well-distributed control and tie points.