# DJI Zenmuse L2 is open for order, please contact LDON for details. #
Zenmuse L2 can be widely used in many different industries as a strong digitalization support. Prior to the product launch, DJI conducted testing and performance validation across major industries. Let's take a closer look at Zenmuse L2's performance in various scenarios, such as terrain mapping, forestry surveys, and data collection.
Terrain Mapping · An Open-Pit Mine in Yunan, China: DJI L2 Data Collection with Accuracy Better Than 5cm Precise measurement results significantly impact mining operations and ecological restoration. Therefore high-precision mine surveying data is essential with accuracy requirement better than 5cm. New Coordinates Technology Co., Ltd. utilized Zenmuse L2 for a surveying project in an open-pit mine site at Yunnan. Over an area of 0.3 square kilometers, the team evenly distributed 215 elevation control points. They set parameters for flight altitude at 160 meters, flight speed at 15m/s, overlap at 50%, and an outward extension of 35 meters. They performed the operation using terrain-following flight, completing field data collection within just 10 minutes. DJI Terra provides support for L2 point cloud data post-processing. Project team conducted precision quality checks on all 215 control points for DEM, DOM, and DLG. The vertical error was determined to be 4.8cm, meeting the accuracy requirements for the project's deliverables. Zenmuse L2 excels expectations in both precision and efficiency no matter it is used in providing rapid and precise surveying guidance for active mining operations or facilitating ecological restoration of abandoned mine.
· Lushan Mountain Scenic Area: L2 Efficiency Doubled Compared to L1, Achieving Tasks with High Efficiency Lushan Mountain Scenic Area has a high vegetation coverage. Guangxi Wanhang Xingkong, conducted surveying operations using L1 and L2 separately within the park, covering an area of 0.5 square kilometers, to acquire 1:500 orthophoto images and topographic maps. With L1, key settings including flight altitude at 100 meters, flight speed at 10m/s, and overlap at 40%, project team completed operation in 26 minutes. In contrast, when utilizing L2 for the operation, the project team adjusted the flight altitude to 150 meters and the flight speed to 15m/s, while maintaining the same overlap rate. L2 completed the task in just 14 minutes, doubling the efficiency of L1, and achieving data accuracy within 5cm.
L2 offers a 30% increase in range compared to L1, with typical operational flight altitudes up to 150 meters. Furthermore, L2 supports 5 echoes, enabling the acquisition of a greater amount of point cloud data within a given time frame. This enhancement further ensures both flight safety and efficiency. Furthermore, due to the undulating terrain and poor signal conditions in the survey area, L2 can leverage the DJI Terra PPK function to match offline base station data, enabling the reconstruction of high-precision models. For the required orthophoto image data, L2 is integrated with a visible light camera, allows project team to directly produce orthophoto images using the photos captured during two flight missions. This eliminates the need for separate data collection, increasing fieldwork efficiency while reducing equipment costs.
Forestry Survey · Guizhou National Reserve Forest Survey: Individual Tree Segmentation and Forest Stock Volume Li Ping County, in the mountainous southeast of Guizhou Province, known as the "Hometown of Chinese Fir," the county has a national reserve forest covering 307,000 hectares with a forest coverage rate of 72.75%. The latest round of national reserve forest inventory work has begun, requiring the survey of 100 hectares of sample plots and the calculation of volume through individual tree segmentation. Previous National Reserve Forest Survey relied on manual work, which were inefficient, labor-intensive, and costly. Li Ping County's National Reserve is very thick with average tree height up to 20 meters. Traditional survey methods were challenging, leading to significant errors in volume calculations. To address these challenges, the Forestry Department delegate Guizhou Yuntu Kanjing Geographical Information Technology Co., Ltd., to use Zenmuse L2 for forestry survey operations. Using M350 RTK equipped with Zenmuse L2, the team collected approximately 1 square kilometer of point cloud data within the National Reserve Forest of Li Ping County. The project team set the flight altitude at 110 meters, flight speed at 9m/s, overlap at 35%, and used 5 echoes, repeat scans, and terrain-following flight for data collection. They completed the fieldwork in two missions, taking a total of 50 minutes. For post-processing, the project team used point cloud colorization and performed individual tree segmentation to extract the number of individual trees and tree heights, which were then used for subsequent tree stem diameter measurement and volume calculation.
To validate the results of L2 operations, the project team compared L2's individual tree segmentation results with high-resolution Digital Orthophoto Maps (DOM) and found that the overall accuracy of tree segmentation exceeded 90%, meeting the expectation of the National Reserve Forest Survey. Additionally, through on-site comprehensive surveys of exceptionally tall trees (exceeding 30 meters) and random checks of regular-height trees (lower than 20 meters), the results indicated that the tree height error extracted from lidar point cloud single-tree segmentation was less than 0.5 meters, also meeting the requirements of the National Reserve Forest Survey.
Zenmuse L2 possesses strong penetration capabilities. When comparing the individual tree segmentation results based on L2 point clouds with high-resolution orthophoto images and manual survey results, it demonstrates higher accuracy in tree segmentation. This capability makes it a powerful tool for National Reserve Forest Inventory Survey.
· Mangroves in Guangxi: Strong L2 Penetration, Subcanopy Point Accuracy at 10cm Mangroves play a crucial role in purifying seawater and maintaining biodiversity. The Shan Kou Mangrove Ecological National Nature Reserve in Guangxi is the second national-level mangrove nature reserve in China, dedicated to protecting the local ecological environment. Orthophoto models is requested for this 100 square kilometers of mangrove forest and site surveys for new incremental mangrove site. Site surveys for mangroves demand highly accurate subcanopy elevation data. Factors such as tree height, inundation height, inundation duration, and exposure time significantly impact mangrove growth, making elevation data with a resolution of less than 10cm critical. While satellite remote sensing technology often offers inadequate resolution, fixed-wing lidar equipment can be expensive. Therefore, the project team M350 RTK equipped with Zenmuse L2 for the operation. Since mangroves grow in intertidal zones, the fieldwork needed to be completed before high tide. The best data collection window for this project was around noon. Given the short timeframe, heavy workload, and complex terrain environment, the project team decided to conduct fieldwork between 3 am and noon.
The personnel set the flight parameters as follows: a flight altitude of 80 meters, a flight speed of 10m/s, and a gimbal angle of -90°. They operated in a dual-echo, repeat scan mode, which allowed them to capture uniform and high-quality point cloud data. Zenmuse L2 offers a 30% increase in range compared to L1, with a spot size of 4 × 12 cm@100 meters, which is only 1/5 of L1's spot size. This enables L2 to penetrate dense mangroves and generate more accurate digital elevation information.
The final point cloud results revealed that in flat areas, the accuracy ranged from 3-5cm, and the ground points beneath the mangroves were evenly distributed, with accuracy exceeding 10cm. This level of precision meets the requirements for a 1:500-scale elevation accuracy.
· Guangzhou Tower 3d model reconstruction
The Guangzhou Tower stands at an impressive 600 meters, making it the tallest tower in China. As part of urban 3D reality project of Guangzhou, a 3D model reconstruction of the Guangzhou Tower is required. But its steel framework makes visible light reconstruction inadequate. Instead, the project team decided to the M350 RTK with the Zenmuse L2 to collect facade data.
The project team used M30 to collect general info of the tower. Based on the rough model, they planned the waypoints and flight path for the L2 operation. The drone flew at 60 meters away from the tower with the gimbal angle at 0° and non-repetitive scanning mode. The entire mission took 3 flights. After flight, DJI TERRA was used to generate LAS point cloud data. The final point cloud data of the tower's structure is intricate and detailed, meeting the project team's requirements.
In the construction of city-level 3D reality, landmark buildings often bring challenges in data collection. Conventional aerial survey methods often struggle to achieve high-quality results. By using Zenmuse L2 to collect LiDAR point cloud data of the building facades, a detailed component-level 3D reconstruction was achieved.
With its high precision, efficiency, integration capabilities, and cost-effectiveness, Zenmuse L2 brings convenience to users across various industries. "Seeing depth, recognizing distances, and striving for perfection" – we believe its introduction marks the beginning of a new era in the toolification of aerial survey Lidar radar. It also represents a significant step forward in DJI's mission to make aerial surveying more accessible to the masses.
Comments