Development And Application Of Key Technologies For Stand Factor Measurement And Statistics

Forest ecosystem is one of the most stable and abundant resource that human beings rely on for survival and development.It plays an irreplaceable role in supporting global ecosystem balance.How to correctly understand and manage forest ecosystem is always the focus and hot issue of scientific research.Forest stand is the smallest regional unit of forest division in forest ecosystem,and its characteristic factor is the basic data reflecting the present situation of forest ecosystem and its change patterns.Therefore,in many fields such as forestry,ecology and environment,stand factor has become the key data of related basic research and is in great demand.How to acquire and statistically analyze these stand factors with high efficiency,high quality and low cost is of great practical significance for correctly understanding and guiding the planning and decision-making of forest ecosystem.In order to achieve this goal,the integrated flow of stand factor measurement,calculation,statistics,prediction and analysis is collectively referred to as stand factor measurement process in this paper.Taking Beijing as the research area,we focus on the core requirements of stand observation,monitoring/detection,modeling and prediction.With"Internet+3S" as the technical support,we have developed key technologies in the field of stand factor measurement,such as stand informatization observation technology,stand micro-sample plot geospatial sampling technology,and stand biomass/carbon reserve dynamic prediction technology.At the same time,the paper discusses and evaluates the realistic demand,theoretical feasibility and technical applicability of the key technologies of stand factor measurement from the aspects of theory,method and practice.The main research results are as follows:(1)Research and development of stand informatization observation technologyWe introduced a handheld PX-80 laser scanner to stand observation in China and verified its measuring ability in flat and mountainous areas respectively.The experimental results show that the evaluation indexes,such as Bias,RMSE,rBias and rRMSE,all meet the precision requirements,and are recommended as a kind of stand informatization observation method.A real-time dynamic multifunctional stereo photogrammetric tree measuring system,RTK dendrometer,has been developed.According to this,the "bidirectional cross photogrammetry" and the "route simulation method" are designed.The measurement and calculation of individual tree parameters such as position of sample trees,DBH,tree height and arbitrary height diameter are realized.The three-dimensional point cloud construction,tree center coordinate position,average DBH of stand and average height of stand are realized.The experimental results show that the terrestrial photogrammetry has the advantages of high efficiency,flexibility and low cost and is expected to have a good application prospect in stand observation.(2)Research and development of stand micro-sample plot geospatial sampling technologyBased on the "Internet+3S" thinking and the principle of geographical and spatial similarity,a homogenization and abstract clustering method of large-scale stand microsample land was proposed,and the whole area of Beijing was clustered into 300?1000 stand micro-sample plot.At the same time,5-9 trees method was used to retest the selected microplots at fixed time,fixed point,positioning and calibration,so as to realize the purpose of dynamic monitoring and detection of stand growth.In order to evaluate its feasibility,100 micro-sample plots were randomly selected,and the extraction results of stand factors were compared with the corresponding standard sample plots.The results show that the sampling precision of the micro-sample plot method is close to that of the standard sample plot method,and the extraction results can represent the overall average situation of a specific stand.On the basis of the micro-sample plot field survey,a 3D point cloud method for measuring the micro-sample plot of stand was constructed by using the RTK dendrometer's surround continuous photogrammetry.The three-dimensional visualization of the micro-sample plot is realized,and the point coordinates,DBH,and tree height can be extracted according to the point cloud model,and then the average DBH of stand,average height of stand,density of trees,stand volume and other stand factors can be calculated.The experimental results show that there is a strong correlation between point cloud micro-sample plot method and field micro-sample plot method,and the extraction precision of stand factor is high.At the same time,the point cloud micro-sampling method also has the advantages of 3D visualization,high data acquisition efficiency and low labor cost.It is a recommended sampling method for stand micro-sample plot.(3)Research and development of stand biomass/carbon reserve dynamic prediction technologyTaking 214 consecutive inventory fixed plots in Beijing as the research object,the time(stand age)parameter was taken as the leading key factor based on the data of three consecutive periods of fixed plots in the sixth,seventh and eighth periods under the premise of considering the influence of environmental factors.With the help of SPSS Modeler software,a multiple regression prediction model of stand biomass based on allometric growth equation was established,and the prediction of stand biomass and carbon reserves in the future time period was realized.The results show that the R2 of both the modeling sample and the test sample are above 0.82,indicating that the fitting degree of the model coefficient was good.The standard error of estimate(SEE),the total relative error(TRE),the mean systematic error(MSE),the mean prediction error(MPE)and the mean percent standard error(MPSE)of the estimated value also meet the accuracy requirements.In addition,the dynamic prediction model of stand biomass based on BP artificial neural network was established by Matlab.For both training samples and test samples,their R2 were above 0.88,and the fitting effect was better than the multiple regression model.The evaluation indexes such as SEE,TRE,MSE,MPE and MPSE also meet the accuracy requirements and are slightly better than the multiple regression model.At the same time,the biomass and carbon reserves of 214 fixed sample plots in the next period were predicted and analyzed by using the established multiple regression model and BP-ANN model respectively,and the stability and applicability of the two models were proved.