| Image is an important information source of human intelligence. It is the primary media of communication. Along with the concept of information highway and digital earth being proposed, the need of image processing technique increase steadily. At the same time, the developments of computer provide a platform for image processing technique.Image processing mainly includes 4 parts: image enhancement, image segmentation, edge detection and image coding. The bionic algorithm presents an effective way to solve these several image processing problems. Basic theories of Genetic Algorithm (GA), Ant Colony Algorithm (ACA), Particle Swarm Optimization (PSO), Fuzzy Logic Theory and Pulse-Coupled Neural Network (PCNN) are illustrated. Then we applicate these bionic algorithms to actual image processing problems. In the image enhancement, the problem of the image enhancement being converted into a problem of nonlinear transformation function, quantum genetic algorithm being applicated in parameters optimization of nonlinear transform function, and we attain favorable result; In the image segmentation, the problem of the image segmentation being converted into a problem of maximum entropy principle, ant colony algorithm being applicated in fitness function optimization based on improved maximum entropy, and we attain very high efficiency; In the edge detection, the Pal algorithm being improved based on paticle swarm optimization, under the new fuzzy edge-detection algorithm, we attain perfect edge location; In the image coding, the image being irregular segmented by pulse-coupled neural network, then the image being reconstructed based on orthogonal basis in each segmentation area, and we attain very good compression.At the same time, the digital image processing system based on TMS320DM642 DSP was presented. We describe the circuits of hardware datailedly. The hardware system is used to detect the proposed algorithms. The experiment results show that the performance of the system is satisfied. |
PDF Full Text Request