Photogrammetric Principles And Techniques For Tree Measurement

Forest resources survey and monitoring aims to identify the quantity and quality of the country’s forest resources, to be informed of their changes, and to provide data to support decision-making on their sustainable development. It is the foundation of forest management of China. The traditional forest survey method used in our country was laborious, inefficient and time-comsuming. As technology advances, more and more forestry workers began using a variety of advanced scientific means to carry out the surveys, especially the fast, non-destructive, non-contact methods that are made possible with the development of photogrammetry and remote sensing technology.This paper concerns the forest farm attached to Huangnihe bureau of forestry in Jilin Province and carries out research around "surveying the forest resources using photogrammetry and remote sensing technique", the important issue of forestry science and technology research. Main results includes analysis of the application and limitation of remote sensing inversion, UAV aerial photography and close-range photogrammetry on forest survey, and proposition of the binocular-camera-single-wood observation and sample field observation method, and analysis of the error sources and factors of accuracy.Firstly, the method of stock volume model inversion based on remote sensing image was introduced. Specifically, this method does the geometric and radiation correction of the selected remote sensing image, extracts its texture and spectral factors, calculates the correlation between the stock volume and the factors that are extracted from the stock volume survey data, combines that information with the alternative factors extracted from correlation analysis of stock volume model, and finally establish multivariate linear model between stock volume and alternative factors by using statistical analysis. Subsequently by making use of the satellite image of Jilin Huangni district taken Resources the 3rd Satellite as the remote sensing data sources along with with small classes ground survey data, The stock forest volume inversion model of Huangni district was established, and its accuracy was analyzed. Meanwhile the limitation of Remote Sensing Inversion method in forest survey was proposed, namely the limitation on the acquisition of remote sensing data and ground data, the limitation on the model, and the limitation of the investigative capacity.Secondly, the UAV aerial photogrammetry method in forest surveys was analyzed, including, from UAV image data through different methods, the extraction of single Tree Height, the average crown canopy, the average height, the canopy density, etc. The inversion model of DBH and volume by extracting factors and field survey data were also established, analyzed the accuracy of the above factor through the UAV flight data and ground survey data, and established the inversion model of local stock volume and DBH.Finally, given that the remote sensing survey and the UAV can get through the inversion method only the DBH data, I tried using terrestrial photogrammetry to do single wood and field investigation, and analyzed the commonly seen problem using current existing photogrammetric methods, namely, only single wood or small scale investigation is feasible given that external auxiliary conditions are constantly required in the computation. Responding to this I proposed the method that use a particular binocular camera of "self-made baseline, rotating tripod shooting, crossing or upright photography" for forest survey methods. This method chooses upright or crossing photography to expand effective observation area, uses a tripod and rotating shooting to obtain the transformation relationship among multiple sets of three-dimensional space, and expand the scope of observation. The software for processing binocular camera was developed, its features including an auxiliary coordinate system relative orientation, the same name point selection, relative coordinates solver, diameter height finding, diameter correction, etc.Binocular camera can be carried out in the relative coordinate system without outside aid conditions in coordinate calculation, and can be directly extracted from a stereo pair of single tree DBH, tree height, volume, trunk type and other factors with an accuracy as high as above 90%; by mounted on a tripod and rotating observation area, it can expand the observation range and get the coordinate conversion relationship between multiple shots, and thus complete the shooting of "circlular photo plot"; it can obtain each target DBH, tree species, the relative positional relationship, and can complete a variety of plots conversion, plots volume calculation, and mixed angular computing by field investigation. Through modeling and experiment I analyzed the binocular camera forest observation error sources and accuracy, finding that the accuracy is correlated with the baseline length B of binocular camera and the shooting distance L. The longer the baseline length, the shorter the distance shooting, the higher the final accuracy. I subsequently obtained the optimum relationship between binocular camera baseline length and photographing distance, i.e. L≤10B. the result of DBH tree height and volume measuring experiment show that the average fractional error were 1.8%,3.6% and 6.1%. In the sample region to establish a photograph plots, the calculated maximum fractional error of stock volume is 8.2%, the average fractional error was 5.5%. In summary, this method effectively shortens the time and improves the efficiency of the investigation, so that the findings can be long-term preserved in the form of photographs, and can also be utilized repeatedly.