The Design And Development Of Tree Ring Analysis System Based On Computer Vision And R Language

Tree-ring studies are one of the most important research objects in many disciplines,including forestry,ecology,and dendrochronology.However,several existing commercial software for tree ring analysis is costly,and some freeware lacks a graphical user interface and is therefore not compatible with some operating systems.These problems have brought difficulties in related fields.With the rapid development of open-source software R and web application development in recent years,advanced computer vision technology has been introduced into tree ring image analysis and therefore the R package used for the measurement of tree-ring parameters and the corresponding cross-platform web application have been designed and implemented.This system is able to extract tree ring parameters and implement statistical analyses in a more precise and effective way,which has significantly improved the level of forestry informatization in China.This paper develops an R package,named MTreeRing for the measurement of tree rings.The package is written in R language that applies many image processing techniques to the automatic detection of tree ring features.This paper also propose two geometric models to correct the errors in tree ring width measurements.To remove obstacles generating by command line interface,a B/S architecture-based Web application for tree-ring analysis has been developed using HTML5,Shiny framework,and MySQL within the R environment,which mainly provides the tree-ring image measurement,crossdating,detrending and chronology building,climate analysis,mapping,data format transformation and data storage.It also supports off-line mode within the local R environment.The system described in this paper satisfies the needs of measuring and analyzing tree rings,especially for users not familiar with R.Users do not need to pay for software.Modular design and the implementation of open source software make the system have considerable potential in secondary development.Our tests suggest that this system is reliable,strong error-tolerance,user-friendly and ease of use.