A First look into the DJI Matrice 4 Enterprise Efficiency and Accuracy

With the recent release of the DJI Matrice 4 Enterprise (M4E), DJI Enterprise has introduced a successor to the Mavic 3 Enterprise, establishing a new standard for precision aerial mapping. This compact yet versatile drone consistently delivers high-accuracy data, making it ideal for demanding surveying applications.

M4E vs. P1 Comparative Test and Analysis

Richard Butkus III, a leading UAV mapping specialist and founder of SURVAIR, has extensive experience in aerial surveying, geospatial data acquisition, and UAV-based mapping solutions. With a background in precision surveying and a passion for advancing drone technology, Richard has been at the forefront of evaluating cutting-edge UAV systems to ensure they meet industry standards for accuracy and efficiency.

With first hands-on access to the DJI Matrice 4 Enterprise (M4E), Richard conducted a thorough assessment of its accuracy and efficiency by comparing it with the industry-standard DJI Matrice 350 RTK paired with the Zenmuse P1 camera. His evaluation focused on key performance factors, including data capture time, data volume, and accuracy results, across both NADIR and oblique imaging scenarios. These metrics are essential in assessing UAV performance for applications such as orthomosaic mapping, surveying, and 3D reality capture, ensuring that professionals can rely on the system for high-precision geospatial data acquisition.

Efficiency Comparison

Since the Zenmuse P1 comes equipped with a default 35mm lens, while the Matrice 4 Enterprise (M4E) features a 24mm lens, this difference could influence efficiency comparisons. To ensure a fair and accurate evaluation, both systems were configured to capture datasets at a similar Ground Sample Distance (GSD) for both oblique and NADIR flights. Additionally, the same Digital Elevation Model (DEM) was used for terrain-following, ensuring consistency in flight altitude and data acquisition across both platforms.

One key takeaway from these test flights is the importance of selecting appropriately sized ground control points (GCPs). With a GSD of approximately 1.5 cm/pixel, the 4-inch (10 cm) GCPs appear at roughly 7×7 pixels or fewer in the captured images, making precise identification and marking challenging. To improve accuracy, larger targets, such as the standard 2' × 2' GCPs, with a high-contrast color pattern, are recommended for better visibility. Alternatively, reducing the GSD (increasing resolution) can enhance target clarity and improve dataset precision.

SURVAIR opted for smaller GCP targets because their P1 system typically operates at a much finer GSD, allowing for the capture of highly detailed models. For these tests, we selected a 1.5 cm/pixel GSD, as it strikes a balance between accuracy and operational efficiency, aligning with most user preferences.

However, when comparing GCP identification after the aerotriangulation process, the P1 system demonstrated a clear advantage. Its larger 4.4-micrometer pixel size provides a noticeably better dynamic range than the 3.3-micrometer pixels of the M4E, enhancing contrast and detail. This makes it easier to identify targets, even when both datasets maintain similar GSD and resolution. In contrast, the smaller pixels of the M4E struggle more to distinguish ground targets from their surroundings, highlighting the benefits of a larger pixel size for photogrammetry applications.

GCP 303 captured by the Matrice 4 Enterprise Camera                                           GCP 303 captured by the Zenmuse P1 Camera

Flight Duration and Data Capture Quantity Comparison

When evaluating flight efficiency, two key criteria are considered:

1. Flight Duration – The total time required to complete the mission, which directly impacts operational efficiency and battery usage.

2. Data Capture Quantity – The number of images captured during the flight, which affects both processing workload and final output quality.

A more efficient system achieves a balance between minimizing flight duration while ensuring sufficient data capture for high-quality mapping and modeling. In this comparison, we analyze the performance of the Matrice 350 RTK with the Zenmuse P1, the Matrice 4 Enterprise (M4E), and the Mavic 3 Enterprise (M3E) to determine which system provides the best balance of flight efficiency and data collection for a typical 40-acre, 1.5 cm GSD mapping mission.

Flight Mission Plans

Below are the detailed NADIR and Smart Oblique flight mission plans for the Mavic 3 Enterprise (M3E), Matrice 4 Enterprise (M4E), and Matrice 350 RTK with Zenmuse P1. Each section includes the mission parameters followed by corresponding flight plan screenshots for better visualization.

NADIR Flight Plans

For NADIR mapping missions, all three drone systems follow the same flight parameters:

Mission Parameters

• Ground Sampling Distance (GSD): 1.5 cm

• Overlap Settings: 70% side overlap, 80% frontal overlap

• Terrain Following: Applied AsterGDEM

Flight Plan Screenshots:

Mavic 3 Enterprise NADIR Flight Plan

Matrice 4 Enterprise NADIR Flight Plan

Matrice 350 RTK + Zenmuse P1 NADIR Flight Plan

Smart Oblique Flight Plans

DJI Smart Oblique is an automated photogrammetry feature that optimizes oblique image capture by adjusting the camera angles mid-flight, reducing redundant images, and improving 3D model accuracy while saving time and storage. The feature was Initially introduced with the Zenmuse P1, As the Smart Oblique feature is now available on the Matrice 4 Enterprise. This raises an interesting question: Can the M4E achieve the same level of efficiency as the P1 system?

For Smart Oblique missions, the following settings are applied across all three drone systems:

Mission Parameters

• Orthophoto GSD: 1.5 cm

• Oblique GSD: 2.12 cm

• Overlap Settings: 70% side overlap, 80% frontal overlap

• Oblique Camera Angle: 45°

Flight Plan Screenshots:

Mavic 3 Enterprise Smart Oblique Flight Plan

Matrice 4 Enterprise Smart Oblique Flight Plan

Matrice 350 RTK + Zenmuse P1 Smart Oblique Flight Plan

The differences between a traditional 5-directional oblique mission and the Smart Oblique mission are significant:

Analysis of Flight Efficiency

The Zenmuse P1 system demonstrates a clear advantage in efficiency due to its high-resolution sensor, which requires fewer photos while significantly reducing total flight duration. Compared to both the M4E and M3E, the P1 system minimizes time on-site while maintaining high-quality data capture.

When evaluating flight duration, the M4E closely matches the performance of the P1, showing significant improvements over the Mavic 3 Enterprise (M3E). Specifically, for a 40-acre, 1.5 cm GSD mission, the M4E achieves a 17-21% reduction in operational time compared to the M3E. This improvement is largely due to enhanced aviation efficiency, making the M4E a notable advancement over its predecessor.

Additionally, enabling the Smart Oblique feature enhances efficiency, saving up to 13% of operational time across all drone platforms. However, this feature increases the number of photos taken for processing, creating a trade-off between reducing time on-site and extending processing time. For conventional 5-directional Smart Oblique missions, the aircraft must fly to the next flight route after completing one, adding transition time and impacting the overall estimated flight duration. Operators who prefer shifting workload to post-processing rather than on-site data capture may find enabling the Smart Oblique feature beneficial. While this feature increases image capture, its impact is less significant for smaller projects, as the area requiring multiple image captures is smaller compared to larger operational sites.

A key consideration is the effect of flight duration near the UAV’s battery limit, such as comparing a 37-minute flight to a 42-minute flight. Although a 5-minute difference may seem minor, it can have a substantial impact due to UAV flight time limitations. For example, the M3E may struggle to complete a 42-minute Smart Oblique mission (24 minutes 11 seconds + 18 minutes 10 seconds) on a single battery, whereas the M4E can confidently complete a 33-minute 24-second Smart Oblique mission without requiring a battery swap. The additional time required for landing, battery replacement, and re-establishing RTK connectivity can significantly extend the total operational duration and increase the risk of user error.

This is not to suggest that these compact drones are incapable of multi-battery operations. However, a longer flight duration, combined with a more efficient aviation system and high-resolution sensor, is more suitable for larger-scale projects. If the operational site is frequently larger than our test scenario, investing in a more advanced system like the Zenmuse P1 can significantly reduce data capture time and enhance overall efficiency.

Accuracy Test

Richard tested the M4E over a 40-acre survey area consisting of a fully developed property with various structures, paved surfaces, and lightly wooded sections. The eastern boundary was marked by a public road, while the northern and western perimeters were densely wooded. A powerline clearing defined the southern boundary.

Ground Control and Check Points

A Spectra SP60 was used to survey and record the coordinates of ground control points (GCPs) and checkpoints (CPs). The coordinates were captured in the NAD83 New Jersey State Plane coordinate system with NAVD88 heights in U.S. feet.

A total of 13 known points were surveyed, with 8 designated as GCPs and 5 used as checkpoints.

Spectra SP60 in operation collecting control point coordinates

Photo of a sample control target used. Targets are 4" (10cm) in diameter

All ground control points (GCPs) and checkpoints (CPs) were later imported into DJI Terra for accuracy assessment. The accuracy report was generated after the aero-triangulation process within the DJI Terra software.

Ground Control Points (8 Total)

Check Points (5 Total)

UAV Camera and RTK Referencing Source

Richard uses two UAV system for this test: M4E with the D-RTK 3 multi-purpose station for RTK referencing; M350 RTK with the Zenmuse P1 with the D-RTK 2 mobile station for RTK referencing. Both systems completed two test scenarios at the same location on the same day. And both D-RTK system were set over on different known coordinates recorded by the same Spectra SP60 GNSS receiver.

Accuracy Flight Test Environment

Pilot: Rich Butkus III

Date and Time of Data Acquisition: December 31, 2024, from 12:00 AM to 4:00 PM EST

Environmental Conditions:

• Partly cloudy

• Maximum Wind Gusts: 7 mph (11.3kph)

• Temperature: 48°F-52°F (9°C-11°C)

Accuracy Test Results

After incorporating all 13 known points into the four datasets and completing the aero-triangulation process within DJI Terra, we reprocessed each flight without applying GCPs to compare accuracy under different conditions.

Richard conducted four flights in total: two for orthomosaic generation and two for smart oblique capture. In the initial processing batch, 8 GCPs were used to constrain all four flights, while the remaining 5 known points served as checkpoints to validate the accuracy of the outputs.

Once all four flights were processed with GCPs, we reprocessed them again—this time without applying any GCPs. Instead, the previously used 8 GCPs were reassigned as checkpoints, increasing the total number of checkpoints to 13. This allowed us to assess the accuracy of both RTK-only processing (without GCPs) and GCP-constrained processing.

Following processing, we extracted the RMSE and median error values from the checkpoints and compiled a comparative analysis chart to evaluate the accuracy of the M4E and P1 systems.

DJI Terra Processed Aerotriangulation Model

Apply GCP and CPs within DJI Terra

DJI Terra Reconstruction Quality Report

Accuracy Test Result

Accuracy Test Results Analysis

From the accuracy test results, several key observations can be made:

1. Minimal Improvement with GCPs (Indicating Higher Intrinsic Precision)

The M4E system exhibits minimal differences between RTK-only and GCP-constrained processing, suggesting that its onboard RTK solution and camera system already deliver highly precise results compared to the P1. For example, in Smart Oblique flights, M4E’s 3D RMSE error improves only slightly from 0.035m to 0.032m, whereas P1 shows a more significant reduction from 0.044m to 0.035m. The smaller improvement in M4E indicates that it achieves a high level of accuracy without requiring external GCP constraints, whereas P1 benefits more from GCP correction.

2. Consistently Lower Median Errors (Indicating Greater Stability)

Median error values, which represent the central tendency of errors, serve as a strong indicator of precision. Across all test scenarios, M4E consistently maintains lower or comparable median errors compared to P1. This is particularly evident in Smart Oblique flights, where M4E's vertical median error is 0.022m with GCPs, compared to P1’s 0.029m. A lower median error suggests that M4E produces more stable and consistent results with fewer outliers, further reinforcing its precision.

3. RTK-Only vs. GCP-Constrained Accuracy

While P1 demonstrates significant improvements with GCPs, indicating a higher reliance on external corrections, M4E maintains stable accuracy even without GCPs. In Nadir flights without GCPs, both M4E and P1 have a 3D RMSE of 0.055m, but after applying GCPs, P1 improves considerably to 0.038m, while M4E shows only a marginal improvement to 0.053m. This further confirms that M4E’s RTK-only solution is already highly precise, whereas P1 requires GCPs to achieve optimal accuracy.

Conclusion

While this test set is limited and does not provide a comprehensive comparison of the two photogrammetry solutions, it offers a valuable first look at the M4E’s performance relative to the high-end P1 system.

Overall, the M4E system demonstrates higher precision, as its accuracy remains stable regardless of whether GCPs are applied. In contrast, while the P1 system achieves higher absolute accuracy when GCPs are used, this suggests a greater reliance on external corrections to optimize results.

For applications where precision and consistency are the main priorities—with minimal variation between RTK-only and GCP-constrained processing—the M4E is the better choice for achieving survey-grade accuracy using RTK alone. Its high-precision camera system allows it to deliver reliable outputs without GCPs for smaller sites or with a reduced number of GCPs on larger projects, ultimately improving operational efficiency. However, if the focus is on absolute accuracy, large-scale surveying, and maximizing efficiency, the P1 system offers a more advanced camera and benefits significantly from GCP integration, allowing it to achieve greater final accuracy after correction.

Summary

The DJI Matrice 4 Enterprise offers a well-balanced solution for aerial surveying, combining efficiency, accuracy, and portability. Its integrated RTK system and optimized imaging ensure survey-grade precision, while advanced flight modes enhance data collection speed. Compared to larger UAV platforms, the M4E simplifies deployment and operation without compromising performance.

For survey teams, portability is just as critical as performance. Richard emphasizes the M4E’s field advantages, noting that larger UAV systems—such as the Matrice 350—can be impractical due to their size and storage requirements. In contrast, the M4E’s compact, lightweight design enables easy transport and rapid deployment, making it an ideal tool for both routine mapping and high-precision survey missions.

"From a surveyor's perspective, our field crews typically operate out of their trucks, which are already fully equipped with essential survey instruments, tools, and supplies. Space within these vehicles is extremely limited, making it impractical to store and transport larger UAV systems such as the DJI Matrice 350 (M350) and its associated battery station. Our crews frequently express concerns about the challenges of handling and maneuvering bulky equipment in the field, as it adds to the logistical burden of their daily operations.
The Mavic 4 Enterprise presents an ideal solution for survey teams seeking a compact, efficient, and reliable UAV for routine data collection. Its lightweight and streamlined design allow it to be easily stored within the confines of a survey truck—potentially fitting under a seat—without occupying valuable cargo space. This level of portability ensures that crews can always have a UAV on hand without compromising their workspace or workflow efficiency. Moreover, the M4E is capable of achieving the same level of survey-grade accuracy as the M350 when paired with a Zenmuse P1 camera. This eliminates the need for larger, more cumbersome drone platforms while still delivering precise, high-resolution photogrammetric data. By leveraging the M4E’s advanced imaging capabilities and real-time kinematic (RTK) corrections, survey crews can confidently collect data that meets industry accuracy standards without the logistical challenges associated with heavier UAV systems.
Additionally, not every surveying project necessitates the use of a large drone. Many jobs, such as quick site inspections, topographic surveys of smaller areas, or supplemental aerial imagery collection, can be effectively completed using the M4E. Its ease of deployment and reduced equipment footprint make it a practical choice for surveyors looking to enhance their capabilities without adding unnecessary complexity to their field operations." - Richard Butkus III, SURVAIR