基于信号盲估计和盲辨识的图象处理方法研究

The main study of this project covers the blind signal estimation, the blind systems identication and the application of image restoration. Combining with the characteristics of the image and some constraint condition added on it, we presented several methods for blind image restoration such that it is more effective and rubost for noise enviroment. Besides, some problems such as the construct of signal model, feature extraction and pattern classification, ect. are also studied in this project.

将高性能的一维信号盲估计和系统盲辨识方法推广到二维图象处理中，并结合图象数据的特点和约束条件，提出更有效的图象盲复原方法，使其对噪声具有鲁棒性，能较好地克服局部极小问题，并具有较广的适用范围，又探索通过适当的数学建模，将信号盲分离技术用于图象的滤波、增强及特征提取等，获得新的图象分析方法。