Building texture acquisition and processing based on an unmanned airship low-altitude aerial survey system.

The traditional oblique cameras cannot self-calibrate to complete the three-dimensional (3D) modeling to meet high mapping accuracy requirements. We propose an oblique combined camera system, called a texture camera, aboard an unmanned airship that can simultaneously obtain a single vertical image and four oblique images of an object. The images acquired by this self-calibrating camera system can be used to produce traditional products on a large scale. Additionally, they can be used to accurately determine the orientation elements of each individual camera, which are essential for 3D modeling. With the methods of building texture extraction and color optimization, the processing of 3D building texture can be completed automatically and quickly. Experimental results show that our strategy can meet the precision requirements of an aerial survey topographic map with a scale of 1:500, and that the building facade textures are complete and clear. Using this texture camera system, 3D building models with a high mapping accuracy and realistic textures can be established.

Boundary image reconstruction based on the nonmonotonic and self-adaptive trust region method for electrical impedance tomography.

The trust region (TR) method is an effective algorithm for image reconstruction of electrical impedance tomography (EIT). Aiming at the drawback of the traditional TR method, an improved method named the nonmonotonic and self-adaptive trust region (NSTR) method is proposed in this paper, in which three kinds of modified techniques are shown that help in improving the computational precision and convergence speed of the algorithm. The comparisons with image reconstruction are carried out between the NSTR, Levenberd-Marquardt (LM) and TR methods. Simulation experiment results indicate that for both the LM and TR methods it is difficult to accurately reconstruct concave and multiple boundaries. However the NSTR method cannot only realize accurate reconstruction, but also provide faster convergence speed. A noise simulation test is carried out and results show that the NSTR method also has strong stability in image reconstruction for EIT. This new method presents a feasible and effective way to research on image reconstruction for EIT.