Study On The Monitoring System Of Invasive Alien Species In Hunan Province

Invasive alien species have become a major threat to China’s ecological security and have caused serious economic losses to the production of agriculture,forestry,livestock,and fishery industries,so carrying out monitoring and prevention and control of invasive species has become a priority for biodiversity conservation.However,invasive alien species are widely and sparsely distributed,and the traditional paper-and-pencil data collection method is timeconsuming and laborious,making monitoring of invasive species difficult and collecting information on epidemics weak.To obtain accurate and rapid invasive species field survey data and improve the dynamic monitoring and early warning capability of invasive alien species,this paper conducted a study on the monitoring system of invasive alien species in Hunan Province,and the main results are as follows.(1)A method for automatic recognition of invasive alien species images based on deep residual networks and migration learning is proposed.Using the Res Net-50 model trained on the Image Net dataset as a pre-trained network model,an invasive alien species image recognition model was constructed by combining field-collected and web-crawled invasive alien species image data through a migration learning approach.Experimental results on the image dataset of major invasive alien plants in Hunan Province showed that the model tested with an accuracy of 91.19%,outperforming other reference models that did not use migration learning.(2)A mobile phone client application for the invasive alien monitoring system was developed.The main functions of the mobile client application include: collecting information on the occurrence of invasive species and image information;automatic identification of invasive species categories based on images;visual display of invasive species distribution and risk analysis;scientific popularisation of invasive species knowledge;establishing work reports to spur staff work and facilitate the management,and establishing a systematic invasive species prevention,control and management mechanism for invasive species in Hunan Province.The mobile phone application improves the efficiency of invasive alien species data collection and can enhance the dynamic monitoring and early warning capability of invasive alien species in Hunan Province.(3)The back-end web server side of the invasive alien species monitoring system has been built.The main functions of the server-side application system include data reception,data storage,data management and data analysis.It can receive the collection data uploaded by clients in real-time and store them in the database in a uniform standard data format;it can query,count,and analyze the collected data,and manage the data dictionary of invasive alien species publicly qualified by the state.In addition,an invasive species image recognition model is deployed in the background to reduce the pressure of image recognition on the mobile site and improve the recognition speed.