《基于CART决策树的多层数据无损检测》为作者：宋甫成最新的研究成果，本论文的主要观点为摘要： 敲击检测是无损检测技术中最有效的方法之一。针对自由装填式药柱界面粘接质量中可能出现的多种类型粘接方式，使用敲击检测技术采集到粘接质量的分层数据。通过对敲击检测分层数据的研究，提出了一种结合CART决策树的航空装备无损检测方法。实验证明基于CART决策树的智能敲击检测技术能够有效检测出自由装填式药柱界面多种类型粘接方式中的脱粘缺陷，避免了传统数字敲击检测中面对多种类型粘接缺陷需要重复设置标准值的局限。 关键词：敲击检测，分层数据，自由装填式药柱界面脱粘，CART决策树 Abstract Coin-tap is one of the most effective methods for non-destructive testing technology。 For the various types of bonding methods that may appear in the bonding quality of the interface of the free-filled grain, the layering data of the bonding quality is collected using the Coin-tap technology。 Through the research on the striking detection layered data, a non-destructive testing method for aviation equipment combined with CART decision tree is proposed。 The experiment proves that the intelligent Coin-tap technology based on the CART decision tree can effectively detect the debonding defects in various types of bonding methods at the interface of the free-filled grain, avoiding the traditional digital knock detection in the face of multiple types of bonding defects The limitation of the standard value needs to be set repeatedly。 Key words Coin-tap, Hierarchical data, Free-packed grain interface debonding, CART decision tree。不知是否符合录用要求，望您批评与指正。