Comparing DJI Zenmuse L1, L2, and the new L3 LiDAR in Dense Forest Environments

A Multi-University Field Evaluation Across Multiple DJI Matrice Platforms  

LiDAR has rapidly become a foundational tool in forestry research and operational forest inventory. As new sensor generations are released, research teams are faced with a practical question: how much improvement do new LiDAR systems actually deliver when operating in real forests under real conditions? 

To explore this question, Candrone coordinated a collaborative field demonstration with researchers from the University of British Columbia and the University of Victoria. Together, the group conducted a controlled comparison of the DJI Zenmuse LiDAR systems across multiple DJI aircraft platforms in dense coastal forest environments in British Columbia. 

This article documents the first phase of that work and shares early observations from the field. Detailed quantitative results are currently being processed by the university teams and will be shared at a later stage. 

Why a Real-World Comparison Was Needed 

Published specifications and controlled test environments provide useful insight into LiDAR performance, but forestry researchers rarely operate in ideal conditions. Canopy density, moisture, fog, terrain, and mission design all influence the quality of LiDAR data. Because of this, the participating institutions wanted to evaluate the sensors in a real operational setting rather than relying solely on technical specifications. 

In particular, the team wanted to understand whether the newest generation of DJI LiDAR sensors delivers meaningful improvements in canopy penetration, ground return capture, and overall data density. 

Project Overview 

The field demonstration took place at Malcolm Knapp Research Forest in British Columbia during winter coastal conditions. The location offered a controlled yet realistic testing environment, allowing multiple research teams to operate side by side. 

Flights were conducted using the DJI Matrice 300, Matrice 350, and Matrice 400 platforms equipped with the Zenmuse L1, L2, and L3 sensors. Coordinated flight missions allowed the teams to evaluate how differences in hardware, scanning patterns, and mission design influenced the resulting datasets. 

Demo Objectives 

The field demonstration was designed to evaluate how each aerial LiDAR and drone generation performs in operational forestry conditions. The primary goals included: 

Designing the Flight Missions 

Approximately ten flight missions were conducted to test real operational variability rather than idealized scenarios. Missions were intentionally designed to vary key parameters that forestry teams regularly adjust in the field. 

Mission variables tested 

One of the key questions guiding the experiment was whether newer LiDAR technology could enable faster data collection without materially reducing data quality. 

Testing in Dense Coastal Forest Conditions 

Two distinct forest units were selected for the flights. The first represented a standard coastal forest stand typical of British Columbia, while the second was an experimental high-density stand cultivated for roughly twenty years to maximize canopy density. This allowed the teams to evaluate sensor performance in particularly challenging environments. 

Weather during the flights included overcast skies and intermittent light fog. At higher scanning frequencies, some LiDAR returns interacted with fog in the upper canopy, slightly affecting data in those layers. Rather than being a limitation, this became an important stress test that reflects real operating conditions in coastal forests. 

Early Observations from the Field 

While the full datasets are still being processed, operators and researchers noted consistent trends during the data collection phase. 

Preliminary field observations 

The new star-shaped scanning pattern appears to contribute to this more consistent coverage. Quantitative comparisons between the L1, L2, and L3 are currently being developed and will be shared in future publications. 

The Role of Flight Parameters 

Beyond sensor comparison, the demonstration also highlighted the importance of mission design. Flight altitude, speed, and scanning frequency all influence the quality and usability of LiDAR datasets. 

Key analysis areas currently underway include: 

Preliminary observations suggest that newer LiDAR systems may provide greater flexibility in mission planning, potentially enabling faster data collection while maintaining forestry-grade datasets. 

Data Processing 

All raw datasets collected during the flights were transferred to the participating university teams for independent processing and evaluation using LiDAR360 and other LAS-compatible platforms. The analysis will examine point cloud quality, ground classification performance, vegetation structure modelling, and dataset consistency across sensor generations. 

Looking Ahead 

This collaboration represents an important step toward understanding how LiDAR technology is evolving for forestry research. Independent field evaluations help research institutions make informed decisions about equipment, workflows, and future study design. Partnerships with drone and LiDAR experts continue to allow forestry teams to test and evaluate systems in real life settings before committing valuable resources to a purchase.   

Book a LiDAR Demo for Your Research Team