Research Live Stumpage Embedded Uitrasonic Nondestructive Testing Of Key Technologies

Ultrasonic detection technology is one of the most widely used method in NDT methods. This method is compared with the traditional method with visual easy to carry, practical and complex work environment, quick positioning, accurate internal damage in vivo identification of stumpage classification. So it is necessary to discuss ultrasonic nondestructive testing technology in stumpage vivo.This thesis based on the research of the principles and methods of ultrasonic NDT live stumpage systems, Principles and positioning algorithm of wireless sensor positioning technology, classification principle and classification methods of support vector machine.Complete the realization of the standing ultrasonic nondestructive testing system in vivo.Simultaneously, ultrasonic standing in the living body when testing time domain characteristic value extraction, processing and analysis, using support vector machine to do the identification of internal damage classification.This paper studies conducted as follows:Research on the ultrasonic NDT live stumpage system implementation. System proposed in the design of the combination of ARM and DSP dual-core processing solutions. This program made good hardware support for complex signal processing. Make full use of the ARM and DSP processing performance, with simple interface. ARM use the DSP host interface communicate with DSP without interrupting normal operation of the DSP. In the procession of software design, adding the wireless sensor localization algorithm. Using RSSI location technology, integration of particle swarm optimization to complete address staff positioning information.Analysis and processing ultrasonic testing in vivo situation stumpage internal damage Damage Identification SVM classification. Performed ultrasound in vivo stumpage propagation time domain features for classification, and then based on support vector machine live stumpage internal damage size recognition classifier in identifying the classification process using cross-validation method, valid for identification classification accuracy for increased.The main conclusion of this study is, when stumpage in vivo ultrasonic nondestructive testing, using the method of support vector machine to analysis the signal data for identification of internal damage is feasible and effective. And after using cross-validation method,it is more accurate to make identification and classification.