目的研究火箭发动机绝热包覆层脱粘检测的自动分类识别方法．方法通过对超声检测探头拾取的回波信号进行分析，提取出作为脱粘识别判断依据的信号特征量，利用似然比方法建立模型，对贴有绝热包覆层的发动机试件进行检测．结果检测结果与试件的预先粘接状态相吻合，提出了利用似然比建立模型的方法．结论利用似然比建立模型的方法进行火箭发动机绝热包覆层的自动分类识别是完全可行的 Aim To study the automatic classification and identification of rocket engine thermal insulation coating debonding detection. Methods By analyzing the echo signals picked up by the ultrasonic testing probe, the signal features extracted based on the judgments of the disbondment identification were extracted. The model of the likelihood ratio method was used to test the engine test pieces with the thermal insulation coating. The result of the test is consistent with the pre-bonding state of the test piece, and a method of using the likelihood ratio to establish the model is proposed. Conclusion It is feasible to use the likelihood ratio model to establish the automatic classification and recognition of rocket engine thermal insulation coating