试件或设备内部缺陷的定量识别已成为红外无损检测与传热反问题领域的一个重要研究方向。针对基于表面红外测温的试件内部多缺陷的识别问题，提出了基于参数转化思想和遗传算法的一种新的定量识别方法，初步研究显示了对试件内部多缺陷识别的良好的适用性。针对该方法，以二维试件内部多缺陷试件为例，从缺陷的位置、形状、数目、网格密度、测温误差等各个方面系统地分析该方法的性能。一系列数值试验验证了该方法的有效性，证明了在各种因素影响下该算法均具有非常优越的性能。

Quantitative identification of inner defects in one test piece is a very important research content in both infrared nondestructive testing and inverse heat conduction problem. Recently, a new method, aiming to identify the multiple defects in one single test piece, and based on parameter transformation and genetic algorithm is presented by this paper, and numerical examples have been given to prove primarily the advantages of the method. In this paper, the performance of the method is systematically studied, in consideration of the position, shape, number of the defects, grid number, the temperature measurement error and etc. A series of numerical experiments prove the effectiveness of the method. Excellent performance of the method is shown though under the influence of these factors.