Study Of Acquiring Tree Measurement Factors By Terrestrial Photogrammetry

Forest resource survey is the basic technique of forestry policy formulation and constantly improve forest potential productivity and it needs a large number of forest measuring instruments to be applied to. Currently the measuring instruments with the backward way to obtain data are single functioned, the slow technology development that restricts the achievement of precision, digitization, automation, information technology and internal and external business integration of forest resources investigation. Meanwhile it is the good time that we should promote the development of national forest measuring instruments by new technology in order to replace the general ones which come from abroad. With the increasing maturity of photogrammetry techniques, computer vision and photogrammetry has brought new development opportunities of the forest measuring instruments. Based on the principle of photogrammetry, we developed the hardware and software system by using digital camera. Total Station plus CCD lens and Mini Smart Station. A detailed operational process within and outside has also been made which aims to make use of photogrammetry to improve the ability and efficiency to get the tree measurement factors and the application in forest measuring instruments and forestry.First based on trees vertical growth characteristics, we propose photographic measurement with a single common digital camera to measure any test plots in up and down directions. By this technology we can measure any period length of the object in the forest to solve the inverse photography baseline, recovery real space photography area ratio between the realized measurements, in order to obtain tree measurement factors on the photograph. This study with the foundation of photogrammetry theory, condition of light and photography of the same name coplanar baseline establish a kind of photography ground measurement taking perpendicular direction as the main direction. To test and analyze the effectiveness of the measurement algorithm and prepare the ground stereo photogrammetry system a software with Microsoft Visual Studio development, platform for the limitations of forestry photogrammetry. From experimental analysis we found that the accuracy of the measurements increases with decrease in field depth distance, with increasing of the diameter& standing baseline considering the impact of camera distortion. So we determine the appropriate measurement range as 2-25m and try to make the plot photo standing in a central location. Five plots are measured and calculated. Experiments demonstrate that the mean absolute error is 0.29cm and the mean relative error is 1.99%by using ordinary digital camera for DBH measurement. For example, using an ordinary digital camera in the vertical direction can directly measure any photogrammetric plant standing in diameter, anywhere in diameter and the visual tree height. Thus by calculation we can obtain the trunk cross-sectional area of the tree standing, Standing Volume, strains number density of plots, average DBH, coordinates and automatic generation of sample map.Second, it is not a perfect way getting relative exterior orientation elements of common digital camera between two shots with SIFT algorithm because the processing speed is slow and accuracy of the results is not high. Then we come up with an idea to eliminate the influence of this aspect. We put the CCD lens above the total station in order to obtain the relative exterior orientation elements of two shots and then through the calculation principle of the space intersection, single image horizontal and vertical angles, we achieve the calculation of image auxiliary space coordinate points. Thus it is true that we achieve the function of measuring tree factors between two images by using Total Station plus CCD lens. Meanwhile, Total Station with ranging function helped us so we can take it as an additional condition to measure flat plot, chest high marking plots and circular plots photogrammetry, in order to achieve the transformation from two-dimensional to three-dimensional coordinate. Experiments demonstrate that the mean absolute error is 0.16cm and the mean relative error is 0.86%based on the photographic measurement plus CCD lens Total Station for DBH measurement. For instance, by obtaining 24 photos from a single image of a round sample, we can directly measure the diameter of any tree standing, anywhere in diameter, tree height and then we can calculate the trunk cross-sectional area of the plant standing, standing volume, the average DBH, coordinates and the production sample map.Again the Total Station is relatively bulky, its manipulation and movement is complicated and can not be accommodated in some plots. The application of it is only on distance and angle measurement, positioning and so on. Taking this into account, we can integrate these functions into CCD lens. So a portable mini multifunctional smart station has been developed based on the image automatic recognition theory, the photogrammetry theory, the similar triangles theory and the triangular function theory which can measure the diameter at breast height and tree height. Simultaneously this device integrates charge coupled device (CCD) image sensor within clinometer, central processing unit, memory, liquid crystal screen, power supply and laser ranging sensor. When measuring, it can acquire 2 values, namely the inclination and the distance from the measuring point to the point needing measuring. Moreover, it can record and save the image information measured. The measurements of forest stand like mean height, volume of wood and tree number density can be realized on the above basis. This device realizes the automatic measurement of the diameter at breast height, tree height and diameter at any part, the basic measurement and other functions through its imbedded running program. The results showed that the absolute average error of the diameter at breast height is 0.37cm, the average relative error for the tree height of 215 standing trees is 2.08%. The single height measured average relative error of 101 standing trees is 1.44%.Finally contrast explanation has been shown about the measurement type and range, inside and outside operational processes, accuracy, advantages and disadvantages among the three kinds of photogrammetry instruments in order to better apply these instruments to forest investigation under different conditions. It also provide methods and technical support for the automation, unmanned, intelligence of forest detection surveys and the realization of photogrammetric technology promotion and application in forest resources investigation.