中文摘要：

介绍了Bayesian Networks（简称BNs）产生“Tuned”模板新方法的基本原理以及BNs法与蚁群行为仿真技术和单纯形法组合的方法。通过实际航空影像的实验结果表明，新方法对纹理影像的识别率是令人满意的，同时还将新方法与遗传算法的结果作了对比，结果表明新方法是很有应用前景的。

英文摘要：

This paper introduces the principle of producing texture ＂ Tuned＂ masks with Bayesian Networks（BNs）. In order to improve the quality of masks, it also introduces the two kinds of combined methods： Bayesian Networks and ant behavior simulation, Bayesian Networks and simplex method. The experiments show that the texture classification using the new approach has satisfying results. In addition, by comparing to genetic algorithm experimental results demonstrate the proposed method is promising in the application.