Farmland topographic mapping based on UAV and LiDAR technology

Abstract: Precision Land Leveling (PLL) is the key technology to build high-standard farmland, and thereby improve unit yield of crop. However, the accurate terrain data of farmland is severely lacking for the massive application of PLL in China. Two means are mainly used to collect land terrain data, including the remote sensing and contact topographic mapping. In the contact topographic mapping, the high accuracy of each points can be achieved using the equipment, such as the total station, but the efficiency and accuracy of interpolating data are still far from satisfactory in the case of large areas. In the remote sensing topographic mapping, the aerial photogrammetric or Light Detecting and Ranging(LiDAR) technologies can be used to cover large areas during one flight mission, but the overall accuracy depends highly on the total number and spatial distribution of selected Ground Control Points(GCPs). Therefore, it is necessary to acquire high-precision land terrain data, and then to establish sophisticated terrain model of farmland for PLL operation in time and cost saving. In this study, an Unmanned Aerial Vehicle(UAV) equipped with the low-cost LiDAR and modules of light-weighted Post-Processing Kinematic Global Positioning System(PPK-GPS) was utilized to acquire the multiple remote sensing data in Kaifeng City, Henan Province, China. Firstly, a mean-filter was used to preprocess the data from LiDAR distance measurement, thereby to synchronize with the PPK-GPS data in an improved accuracy. Secondly, the time-varying attitude (pitch and roll) data in the UAV was considered to correct the preprocessed LiDAR distance measurements, in order to remove the impacts of constantly changing attitude of UAV platform. Then, the PPK-GPS data and the corrected LiDAR distance measurements were integrated to calculate the altitude of land surface, and finally 1 417 sets of effective discrete terrain surveying points were obtained in 3-dimensional coordinates (longitude, latitude, altitude). An accurate evaluation was also carried out in the PPK-GPS modules and LiDAR device still for 10 mins consecutively, where the vertical accuracy of positioning data from PPK-GPS and the distance measuring accuracy of LiDAR were 3 cm and 1 cm, respectively. Furthermore, the accuracy of 3-dimensional coordinates of discrete terrain surveying points was validated in the UAV-LiDAR topographic mapping system using a hand-hold PPK-GPS device, suitable for the centimeter-level accuracy. The Root Mean Square Error(RMSE) of terrain surveying method in between the UAV-LiDAR topographic mapping system and the hand-hold PPK-GPS was 5.2 cm, indicating a high accuracy in the terrain surveying domain. A Kriging interpolation model was used to establish a Digital Terrain Model(DTM), according to the 1 417 sets of 3-dimensional coordinates of effective discrete terrain surveying points. The DTM accuracy was slightly lower than that of UAV-LiDAR-based farmland topographic mapping system. This difference can be attributed to the measuring accuracy of individual points, and the spatial distribution of measured points in a ground survey. Therefore, the proposed approach can be expected to improve the efficiency and time/cost saving, while the accuracy of farmland terrain surveying, and thereby to promote high-standard farmland.