-
[show abstract]
[hide abstract]
ABSTRACT: For time-variation and nonlinearity of vehicle image classifier design in complicated environment, a novel classifier based on Immune Clone Concentration Clustering Algorithm (ICCCA) was proposed, inspired by the clone selection theory in the natural immune system. It overcame the defects of traditional classification algorithms as multi-restraint and trap of local optimum. It also avoided the high complexity and regulating dissimilation caused by the cross and variation rules of exiting immune clone clustering algorithms. Directing towards two-class image pattern recognition problem, we defined the antibody and antigen of vehicle image, established class divisibility standard, and designed the affine mathematical model based on the antibody and antigen's Euclidean distance and their concentration. We proposed a method to create immune clone memory cell by the antibodies concentration as long as concentration regulating. The algorithm can guarantee the antibody diversity, and obtain the global optimal solution quickly and precisely. The experimental results verified that the classification accuracy of the algorithm is superior to other classification algorithm.
Natural Computation, 2008. ICNC '08. Fourth International Conference on; 11/2008
-
[show abstract]
[hide abstract]
ABSTRACT: A new Hough-based algorithm termed segment Hough transform (SHT) is proposed to detect the curve in the binary image. The segment-transform, which is the main idea of the novel approach, makes the curve detection very efficient. Thanks to the chain code termed angle chain code (ACC) as well as the modified direction measurement, the segmentation operation is also considerably fast and precise. Compared with the conventional Hough transform (CHT), our algorithm is dramatically faster (by over 20 times) And when the background is complex, the SHT is about four times faster than the random Hough transform (RHT), Furthermore, the proposed algorithm inherits the high robustness from CHT. In some situations, it is even more accurate. We implement the novel algorithm in an embedded system to estimate passengers flow on the bus. The detection rate is over 95% which indicates it is quite suitable for real-time application.
Image and Graphics, 2007. ICIG 2007. Fourth International Conference on; 09/2007
-
[show abstract]
[hide abstract]
ABSTRACT: In this paper, an integrated algorithm to detect bridge objects over rivers was introduced for satellite imagery interpretation. It is composed of two steps: first, segment the river from complex background using data driven strategy; second, detect the bridges in the shrink searching area using knowledge driven strategy. Considering the ubiety of the bridges and the surroundings, this paper focused on the water recognition based on feature extraction. By analogy with the lowest combinative energy according to the biological principle, we proposed brand new criteria which discussed the interrelationship of features. The assess method is adopted in the feature extraction of object image and the optimal combination of water body features is achieved. 50 real satellite images have been tested. Experiment results showed that the method improved the water recognition and therefore increased effectiveness of bridge detection. I.
Industrial Electronics and Applications, 2007. ICIEA 2007. 2nd IEEE Conference on; 06/2007
-
[show abstract]
[hide abstract]
ABSTRACT: A novel chain code and a new clustering-based algorithm are proposed to detect the curve in the binary image. The novel chain code, which is termed angle chain code (ACC), is more proficient and precise in describing the local-shape of the edge than classic chain codes. Thanks to the ACC and the extraction of geometries, the clustering-based algorithm can detect the mathematic models of contours efficiently. Compared with the standard Hough transform (SHT), our algorithm is much faster (by over 690%) and requiring much less memory space. Furthermore, it inherits the high robustness form classic clustering-based approaches. In some situations, it is even more accurate. We implement the novel algorithm in an embedded system to estimate passengers flow on the bus. The detection rate is over 95% which indicates it is quite suitable for real-time application.
Industrial Electronics and Applications, 2007. ICIEA 2007. 2nd IEEE Conference on; 06/2007
