ESTIMATING THE FLYING TRAJECTORY OF TILT UAV PHOTOGRAPHY FROM IMAGE SPACE TO OBJECT SPACE USING IMAGE FEATURES

Hits: 113
Research areas:
Year:
2017
Type of Publication:
Article
Keywords:
Off-nadir photography, Space resection, Exterior orientations, Flying path
Authors:
Min-Lung Cheng, Mass Matsuoka Toshiaki Satoh
Abstract:
To analyze and visualize the environment, optical imagery is one of the great resources to acquire spatial information. In past years, two techniques have been well developed in processing the spatial data, namely photogrammetry and computer vision. No matter which method is adopted, three elements are defined, including the exterior orientation parameters of the camera, the photography, and the ground objects. As for exterior orientations, they can be measured by GPS/INS or estimated via spatial resection with ground control points. However, the optimized space resection may not be able to find the camera exterior orientations due to inappropriate initial values especially for off-nadir imagery. Inorder to overcome this obstacle, this study aims to utilize the relative translations in image space and transform them into object space by image features, such that initial exterior orientations can be approximated. The result indicates that incorrect matched features would lead to great error prorogation for a sequence of photography with the increased flying path. Hence, a few of ground control points are introduced to improve the accuracy and precision of the exterior orientation parameters. As a consequent, this research provides a novel perspective to approximate the initial camera exterior orientation parameters from image space to object space, and conquers one of the important difficulties toward oblique photography when performing the robust spatial data processing.
Full text: 159.pdf [Bibtex]
You are here: Home ACRS ACRS Overview Proceedings estimating-the-flying-trajectory-of-tilt-uav-photography-from-image-space-to-object-space-using-image-features