Search

Forest biomass is an important influencing factor in the study of global environmental and climate change, and plays an important role in the study of the carbon cycle. It is important to estimate forest biomass in a region quickly and accurately. In order to explore the biomass research method at the single-wood scale in the urban forestry demonstration base of Northeast Forestry University (NEFU), the biomass estimation model of the same tree species in the target sample site was obtained, and the above-ground biomass of the target tree species was obtained. In this paper, the Unmanned Aerial Vehicle (UAV) was used to acquire remote sensing images of Pinus sylvestris var. mongolica in the modeling and validation sample plots in the urban forestry demonstration base of NEFU. The UAV remote sensing images were pre-processed and the diameter at breast height (DBH) of all Pinus sylvestris var. mongolica in the modeling and validation plots (Plot1 and Plot2) was measured. The images were segmented and crown information was extracted. The eCognition software was used to extract the information from the remote sensing images and to automatically segment and classify the single tree crowns. The information was imported into ArcGIS software and the crown area of all the Pinus sylvestris var. mongolica was counted. The above-ground biomass of Pinus sylvestris var. mongolica in the sample plots was estimated. Combined with the measured DBH data, the model was fitted to the Pinus sylvestris var. mongolica data in the sample plots, and an optimal model of crown areaDBH was fitted. With the help of the empirical model of DBH - biomass, the above-ground biomass of Pinus sylvestris var. mongolica was derived. The results showed that the above-ground biomass of Pinus sylvestris var. mongolica in the target sample plots can be effectively estimated by using UAV remote sensing technology and the fitted optimal model. The total biomass of the modeling plot1 was 6 315 kg and the validation plot2 was 5 802 kg. [ABSTRACT FROM AUTHOR]

In order to improve the correctness and efficiency of tree species recognition, to slow down the occurrence of overfitting, and to enhance the practicality of tree recognition techniques, this paper proposed a tree recognition method based on an improved convolutional neural network. The approach started with compressing the channel in the Xception framework to further exploit the average pool of global feature mapping, and changing the hybrid attention connection to parallel connection. Next, a randomly selected part of the feature map was normalized and re-entered into the neural network after using the attentional feature map cropping method. Finally, ablation experiment was performed and the accuracy of tree species recognition was up to 98.90% in the tree dataset when the learning rate was 0.1 and the iteration was 50 times. The study showed that the proposed improved convolutional neural network had better recognition effect on tree recognition. And the memory of the convolutional neural network architecture was reduced to 133.9 MB and the time consumed was only 458 ms. Using the improved convolutional neural network not only improved the accuracy of tree recognition, but also reduced the time cost. [ABSTRACT FROM AUTHOR]