Recognition of space image target based on improved fractal technique
- Paper number
IAC-06-B6.P.1.07
- Author
Dr. Balin Tian, Northwestern Polytechnical University, China
- Coauthor
Prof. Jianping Yuan, Northwestern Polytechnical University, China
- Year
2006
- Abstract
Most of the classical gray features or geometry features were based on Euclidian geometry models, which cannot completely describe the complexity and irregularity of the environment. At present, fractal is used to describe such kind of objects involving irregularity and complex self-similarity. It has been recognized as having great potential in the field of target recognition and image processing. The fractal texture can describe the natural background characteristic information of image correctly. According to the fractal feature of man-made object in space, the object can be partitioned off background based on the difference of intrinsic fractal features between them. In fact, the accuracy of recognition depending on fixed dimension fractal is limited. In order to verify whether space target exists or not by computing form the image and improve the accuracy of object recognition, a new algorithm for space image object recognition based on improved fractal technique is presented in this paper. The core of this method is based on variable dimension fractal theory, that the dimension is dynamic and variational. It is more flexible than fixed dimension. The variable fractal dimension and the error of fractal feature are used to detect man-made object in space background, then we can partition and recognize the space targets. The validity of the proposed method is examined by processing and analyzing images of space targets. The results for real images indicate that the method provides valuable solutions with high precision and a low computational cost, and it is efficient approach for space target recognition.
- Abstract document
- Manuscript document
IAC-06-B6.P.1.07.pdf (🔒 authorized access only).
To get the manuscript, please contact IAF Secretariat.