Zhao Danpei

Full-viewpoint 3D Space Object Recognition Based on Kernel Locality Preserving Projections

Abstract Space object recognition plays an important role in spatial exploitation and surveillance, followed by two main problems: lacking of data and drastic changes in viewpoints. In this article, firstly, we build a three-dimensional (3D) satellites dataset named BUAA Satellite Image Dataset (BUAA-SID 1.0) to supply data for 3D space object research. Then, based on the dataset, we propose to recognize full-viewpoint 3D space objects based on kernel locality preserving projections (KLPP). To obtain more accurate and separable description of the objects, firstly, we build feature vectors employing moment invariants, Fourier descriptors, region covariance and histogram of oriented gradients. Then, we map the features into kernel space followed by dimensionality reduction using KLPP to obtain the submanifold of the features. At last, k -nearest neighbor ( k NN) is used to accomplish the classification. Experimental results show that the proposed approach is more appropriate for space object recognition mainly considering changes of viewpoints. Encouraging recognition rate could be obtained based on images in BUAA-SID 1.0, and the highest recognition result could achieve 95.87%.

Automatic Extraction and Matching of Control Points for Distortion Correction in Star-Background Image

Usually a typical geometry distortion will occur when images are captured because of inaccuracy of axis distance of lens in the optical camera lens. The problem is also obvious in star-background images, which necessitate distortion correction for subsequent analysis. In this paper, a new method based on star- point matching is proposed to extract and match automatically control points in star-background images. We acquire automatically control point pairs using point matching between points in pre-corrected image and points in ideal image which relies on the catalog. This work extends applied domain of Hausdorff Distance (HD) which is one of commonly used measures for object matching. In our experiments, least Trimmed Square HD (LTS-HD) was used in point matching, and the result is effective.

High-speed Recognition Algorithm Based on BRISK and Saliency Detection for Aerial Images

A fast ground object recognition method for aerial images, taking airports, oil depots, harbors etc. as research objects, is proposed in this paper based on BRISK and the visual saliency detection. According to the characteristics of aerial images, such as high resolution and complex background interference, saliency detection is applied to select the candidate object region where the target may exist. Therefore, it can reduce the searching range effectively. And then, BRISK matching method is used to recognize the object efficiently. A variety of experiments under different interference factors are carried out base on the typical object database of aerial images in this paper. Experimental results show that the proposed algorithm can not only maintain the validity of BRISK features under the conditions of rotation, scale, illumination and viewpoint changes, but also shorten the matching time, satisfying real-time demand.