| In Forest Resource Inventory, the stand factors are investigated in order to understand the quantity and quality of forest resources, so as to make forestry planning, designing and operating. Forest Resource Inventory is a time-consuming job, traditional measurement methods require different instruments, and it is also limited by terrain conditions. The common diameter measuring method is diameter tape, but it can’t measure the upper dimeter.It can hardly be used in the complex environment too. The height measurement device is bruce height finder, which need to measure the horizontal distance and it is limited by topographic conditions.For the fast and accurate measurement of trees, we designed and developed a portable digital intelligent dendrometer. This instrument consists of STM32F103VET6 microcontroller, ov9655 sensor, LIS331DLTR angle sensor, and keyboard, laser ranging sensors, monitors, power supply and the relevant circuit module. The ranging uses phase principle, with a ranging accuracy of 0.047m. In the mode of simultaneous measurement of diameter and height, the diameter of the tree can automatically be identified through image processing with camera and CMOS sensor. In addition, there are caliper mode, altimeter mode, distance measurement mode, angle measurement mode and the area calculation mode. The codes compiled by C is written into MCU and each measurement is realized. Volume computation, tree distribution image and comparing index computation are realized n coordination with post processing software. It integrates mapping for field and indoor work.Experiments in Olympic Forest Park and Jiufeng in Beijing as well as in Pengkou, Fujian are conducted.The main contents are as follows:(1) Automatic recognition and measurement of tree diameterImage was taken by camera and processed by CMOS. The image was smoothed by Gaussian inverse filter after conversion from RGB to greyscale. Then edge was detected through non-maximum suppression and double thresholds edge connection. Tree trunk automatically detection algorithm was developed on the base of the detecting image. Tree trunk diameter was computed according to the relationship among focal length, object distance, image distance and pixel size. We validated the image recognition results by selecting different trees. Results showed that the image recognition precision was 87.9% and most data conformed to the requirement of forest inventory. Manual adjustment function is added to adjust the situation whose recognition result isn’t well so as to increase the accuracy of tree measurement. Mini dendrometer can measure the diameter at given height, which is realized by laser finder sensor, inclinometer and CMOS sensor. The verification in the upper diameter measurement showed that:for a tree with a height more than 20m, the error from Mini dendrometer increases with the tree height. In the height of 1-5m, there is a minimum error of 0.37cm. In the height of 5-10m,10-15m and 15-20m, the errors are 0.56cm,0.58cm and 0.58cm, respectively. When the height is above 20m, the error is 0.83m. To analyses from the distance, there is conclusion that the smaller error appears in the last two distance sections. When the tree height is under 13m, the minimum error by the Mini dendrometer in 1-2m,2-3m,3-5m,5-8m and 8-13m are 0.15cm,0.22cm,0.46cm,0.54cm and 0.57cm. For the tree whose upper part is blocked, the diameter could be obtained using the taper equation, with only a diameter at breast height. The taper equations are fitted basing on the models that are proposed by Zeng Weisheng and Mei. We found that the equation based on Zeng’s model has higher precision with a relative error of 9%. The other equation based on Mei’s has a larger error.(2) Fast and convenient measurement of tree heightHeight measurement is realized by using laser finder sensor and inclinometer. The tree measurement of CMOS sensor is realized by means of the transformation from RGB to HIS, tree trunk thining, projection transformation, projection reconstruction and so on.In height measurement, the diameter at breast height has a maximum error of 0.012m, a mean error of 1.298×10-3 m. The measurements heights measured in different angle showed that when the angle is between 35°-60°, the height error is almost within 5%. The error is relatively large when the angle is larger than 60° or smaller than 35°. The error mean relative error is 10.43% when the angle is above 60°, while the relative error is 2.53% and 5.65% in the case of 35°-60°and 0°-35°. In conclusion, the height can be measured precisely during 35°-60°.(3) the measurement of stand factors and forest landMini dendrometer can also measure crown diameter, standing tree volume, canopy density and so on. Taking Mini dendrometer and taper measurements for comparison, the maximum absolute difference of crown width is 0.494m, with a RMSE of 0.283m and a precise of 96.6%. the stand volume of 82 trees are measured. The maximum error between data measured by Mini dendrometer and taper is 13.34%. For the digital angle gauge measurement, the precision is high in the panted forest. The most precise relative error is 1.7%, and the next most precise is 3.7%. The relative errors of other plots are above 10%. However, natural forest is more complex and the relative errors are between 20% and 40%, which is relatively high. The maximum resolute error of volume measured by digital angle gauge measurement is 19.02 m and the minimum resolute error is 1.19m3. |
PDF Full Text Request