Abstract: This article introduces the design idea of the realization procedure of the unknown instruction set recognizer.A key issue is classifying the characteristic standards.By taking a research on a number of binary，assembly codes and disassembly codes，based on the statistic classifying method，we extract the integrity characteristic that is embodied in the code structure and the regularity characteristic that appear in the programming style，accordingly forming tree-type feature recognition system.

摘要： 介绍了指令集识别器的设计思想与实现流程。通过对大量二进制目标代码、汇编源代码和反汇编代码的统计研究，基于统计分类方法，提取代码结构中所体现的整体性特征和编程风格中所体现的规律性特征，形成了树形的识别特征指标体系。