Construction of Digital Surface Model by Multi-sensor Integration from an Unmanned Helicopter

Three dimensional data is in great demand for the various applications. In order to represent 3D space in details, it is indispensable to acquire 3D shape and texture together efficiently. However, there still lack a reliable, quick, cheap and handy method of acquiring three dimensional data of objects at higher resolution and accuracy in outdoor and moving environments. In this research, we propose a combination of a digital camera and a small (cheap) laser scanner with inexpensive IMU and GPS for an unmanned helicopter. Direct geo-referencing is achieved automatically using all the sensors without any ground control points. After the accurate trajectory of the platform with attitude changes are determined through the direct geo-referencing, 3D shape of objects is determined by laser scanner as 3D point cloud data, while texture is acquired by CCD sensor from the same platform simultaneously. A method of direct geo-referencing of range data and CCD images by integrating multi sensors for constructing digital surface model are focused. While measuring, an unmanned helicopter is continuously changing its position and attitude with respect to time. For direct geo-referencing, IMU measures the movement of the platform. IMU has a rising quality, but it is still affected by systematic errors. Through Kalman filter operation, an optimal estimate of the sensor position and attitude are determined from GPS and IMU. Meanwhile, geo-referencing of CCD image is determined by bundle block adjustment. GPS and IMU allow automatic setting of tie-points and they reduce the number of tie-points and searching time of tie-points by limiting of searching area. The result of bundle block adjustment aids Kalman filter. IMU are initialized for Kalman filter using the result of bundle block adjustment. That is, after every bundle block adjustment, IMU and its error are complemented. Geo-referencing of laser range data is done by using the result of aiding Kalman filter. Therefore, geo-referencing of range data and CCD images is done directly to overlap exactly with high accuracy and high resolution. The method of data acquisition and digital surface modelling is developed by direct geo-referencing of laser range data and CCD images with GPS and IMU. This is the way of rendering objects with rich shape and detailed texture automatically. A new method of direct geo-referencing by the combination of bundle block adjustment and Kalman filter is proposed. * Corresponding author. Masahiko Nagai, Tel:+81-3-5452-6417 E-mail: [email protected]