Diagnosis Method Of Standing Wood Moisture Content Based On Multi-sensor Information Fusion

In recent years,extreme drought and high temperature events have increased significantly,resulting in frequent forest decay and tree death on a global scale.The research on the mechanism of plant drought death has become a hot spot,in which the water physiology and water content state analysis of woody plants are the core content.At the same time,modern forestry research and digital management put forward new requirements for the collection technology of forest resources information,especially the real-time and accurate remote automatic collection of tree physiological information.At present,the traditional remote sensing method is too time-consuming and difficult to measure the physiological state of trees,but it can not meet the needs of the traditional remote sensing method.Therefore,aiming at the development needs of tree water physiology research and forestry information technology,this paper proposes to carry out the research on the diagnosis method of standing tree core water content based on multi-sensor information fusion.The main ideas are as follows: firstly,the surface moisture content of standing tree trunk is retrieved through wireless acoustic emission sensor nodes;Secondly,the distribution of soil moisture content in the roots of standing trees was detected by soil moisture sensor nodes;Finally,based on the distribution characteristics of water content on the surface of trunk and root,the water content in tree core is deduced.The specific research scheme is as follows:(1)Based on STM32F103 chip,the soil moisture sensor node is designed and manufactured.The capacitive sensor is used to collect the soil moisture of standing tree roots,and the DHT11 sensor is used to collect the air temperature and humidity data.The relevant data is sent to the host computer through sx1278 chip.At the same time,the host computer interface is designed with pyqt5 to complete the remote real-time collection of soil moisture and other data at the root of standing trees.(2)The wireless acoustic emission node based on stm32f405 chip is used to remotely collect the acoustic emission data of the surface layer of the standing tree trunk,and the acoustic emission feature subset that can best characterize the surface layer of the standing tree trunk is selected through the mRMR feature selection algorithm,which verifies the feasibility of retrieving the water content of the surface layer of the standing tree trunk from the acoustic emission data.(3)The data fusion algorithm of tree trunk moisture content and tree root moisture content is proposed.At the same time,a new SVM classification model is proposed and trained.Sparrow search algorithm(SSA)is used to optimize the(C,γ)parameters.The results show that the test accuracy of the model training set is as high as 96.4%.(4)The trained diagnosis model of standing tree core moisture content is used to measure different coniferous and broad-leaved trees in Nanlin campus.The diagnosis results show that the measurement accuracy can reach more than 95%,which reflects the strong generalization ability and robustness of the model.At the same time,the SSA-SVM model is compared with other classifiers,such as naive Bayes and random forest.The results show that the recognition accuracy of SSA-SVM model is higher than that of other existing model classification algorithms.