UAV Remote Sensing Technology To Estimate The Research Of Eucalyptus Volume

After the traditional aerial remote sensing technology and space remote sensing technology, the UAV remote sensing technology is developing rapidly to be the third generation of remote sensing technology. It is advantage with light and flexible, quick, low cost and repeatedly used, etc. And it becomes a powerful complement of traditional aviation remote sensing and remote sensing. In this article, Nanning Guangxi YuanTanLi management area which is the study area, the high-resolution remote sensing images in the study area are obtained by EBEE UAV equipped with a digital camera. The processed orthogonal image is that can be used to extract a single tree crown width, forest stand mean crown, canopy density, forest stand parameters factor, etc. Extracted single tree crown can be established the DBH - crown regression model with fieldwork DBH, the relational model can also be combined with a single timber volume and DBH to get aviation forest timber volume model, and it is used for the rapid estimation to get the forest stand volume. This study aimed to explore the UAV remote sensing technology in forest resources survey to assist in the investigation of the traditional model, to improve the efficiency of investigation, to provide technical reference for the construction of digital precision forestry. The main conclusions from this study are:1、High the DOM images positioning accuracy.Using total station to measure the coordinates of the study area GCP point on the DOM images for accuracy test, among 39 coordinates of the point, midpoint position of maximum error are 9.3384 pixels, the smallest are 0.3909 pixels, as the average error value are 2.7044 pixels, and the point error of the mean is 0.2199m.2、Small class divisions of DOM images with a high precision.Indoor small class divisions and field investigation are both on the basis of small class divisions DOM images, both methods can get 68 small classes. The indoor small class division is compared with the field investigation, uniformity is 97.89%. The result is compared with the GPS small class areas, uniformity is 98.41%.3、Using DOM images can be extracted a single tree crown width, forest stand mean crown, canopy density, forest stand parameters factor, etc.The method of object-oriented multi-scale image segmentation for images choosing the right combination of scale and segmentation parameters that can extract single tree crown which extracts wood crown single precision of nearly eucalyptus mature forest is 74.09%,70.48% is young forest; and manual interpretation of wood crown single precision of nearly eucalyptus mature forest is 80.32%, On this basis, the average crown can be extracted forest stand mean crown, canopy density, etc. after the manual interpretation method estimates to obtain the accuracy of average crown stands is 87.96%, for the split method is 86.27%.4、The standing timber volume model aviation is established by individual tree crown extraction and field survey data that can be used to quickly estimate the stand volume.Using the method of object-oriented image segmentation to get tree crown and field surveys forest DBH to establish DBH - crown model, the maximum model determines coefficient is 0.679, while the relationship between the diameter and the single volume to establish the regression model, the maximum coefficient of determination is 0.957. It can also set up a dollar volume crown aviation model, while using the modeling outside of 337 fast-growing eucalyptus woods near mature secondary data, the accuracy of the model can be inversion, the average precision can reach 83.61%. And on the basis of the estimated volume of samples, the most accurate is 98.94%, the lowest accuracy is 69.64%, the overall accuracy is 91.20%.