Forest spatial structures are the foundations of the structure and function of forest ecosystems. Quantitative descriptions and analyses of forest spatial structure have recently become common tools for digitalized forest management. Therefore, the accuracy and intelligence of acquiring forest spatial structure information are of great significance.
In this study, we developed a forest measurement system using a mobile phone. Through this system, the following tree measurements can be achieved: (1) point cloud of tree and chest diameter circle to measure tree diameter at breast height (DBH) and position coordinates of tree by using simultaneous localization and mapping (SLAM) technology, (2) virtual boundary creation of the sample plot, and the auxiliary measurement function of tree with the augmented reality (AR) interactive module, and (3) position coordinates and single-tree volume factor to calculate the spatial structural parameters of the forest (e.g., Mingling degree, Dominance index, Uniform angle index, and Crowdedness index).The system was tested in three 32 x 32 martificial forest plots.
The average DBH estimations showed BIAS of -0.47 to 0.45 cm and RMSEs of 0.57 to 0.95 cm. Its accuracy level met the requirements of forestry sample surveys. The tree position estimates for the three plots had relatively small RMSEs with 0.17 to 0.22 m on the x-axis and 0.16 to 0.26 m on the y-axis. The spatial structural parameters were as follows: the mingling degree of plot 1 was 0.32, and the overall mixing degree of tree species was low. The trees in plots 2 and 3 were all single species, and the mixing degree of both plots was 0. The dominance index of the three plots was 0.56, 0.51, and 0.51, indicating that the competitive advantage of the whole orest species was not obvious. The uniform angle index of the three plots was 0.55, 0.59, and 0.61, indicating that the positions of trees in the three plots were randomly distributed. The crowdedness index of plot 1 was 1.03, indicating that the degree of aggregation of the trees was low and showed a random distribution trend. The crowdedness index of the other plots were 1.36 and 1.40, indicating that the trees in the plots show a trend of uniform distribution, and the uniformity of plot 3 is higher than that of plot 2, but the overall uniformity is relatively weak.
The findings of this study provide support for the optimization of forest structures and improve our conceptual understanding of forest community succession and restoration, in addition to the informatization and precision of forest spatial structure surveys.