The Zenmuse P1 is a camera dedicated to photogrammetry. The sacrosanct full-format sensor is finally available at Dji. 36 * 24mm with 45 MP, finally a sensor capable of competing with the best manufacturers on the market.
All its specs are designed for photogrammetry:
- 45MP to capture maximum detail
- Sensor size (full frame) allowing large photosites (pixels) of 4.4um, ensuring minimal noise, wide dynamic range and perfect sharpness.
- A mechanical global shutter, eliminating motion blur for optimal results
The Zenmuse P1 can accommodate a choice of 3 lenses dedicated to this camera and created by DJI: a 24mm, a 35mm and a 50mm. These lenses, also designed for photogrammetry, offer minimum image distortion for optimum precision.
With TimeSync2.0, precision is guaranteed. This module makes it possible to synchronize the capture with the drone orientation information (IMU), the orientation information of the camera and the information of the GPS position. All this information makes it possible to achieve precision in the geolocation of the pictures to the nearest centimeter.
Dji announces a precision of the results of 3cm in 2D, 5cm in 3D for a GSD of 3cm (flight at 165m!).
If you are wondering what is the advantage of such a sensor over a Phantom 4 RTK (which by the way remains an excellent drone for photogrammetry), here is an example:
Imagine that we have a plot of 50Ha to survey and that we must respect a GSD of 1cm maximum.
The Phantom 4 RTK and its 1 inch 20MP sensor will be able to scan 5Ha during a flight of just under 30 minutes, flying at 36m (GSD = 1cm) with recovery rates of 80% and 70%.
The M300 RTK with the P1 and the 24mm lens will be able to scan 50Ha while flying at 50 meters in height (legal maximum in S2 with this type of drone) in order to have a GSD of 0.9cm.
It will take 5 hours for the Phantom 4 RTK, or 10 flights at least to survey such an area. It will take about 40 minutes to raise 50Ha with the P1 mounted on a Matrice 300 RTK, or only one flight! And with fewer photos taken by the P1 compared to the Phantom 4 RTK, the photogrammetric processing will also be much faster.
Another example: the Dji Matrice 300 with its P1 camera is able to model buildings with a resolution of the order of a millimeter. For auscultation of a building for the purpose of looking for cracks or defects of 1mm, you would have to fly about 3.6m from it to distinguish them with a Phantom 4 RTK. With the P1 and the 50mm lens, it is possible to stay 12 meters from this building, so the flight is safer and it is even possible while still approaching to have a GSD less than 1mm. All without noise thanks to its Full frame sensor.
The Matrice 300 RTK + P1 combo will therefore be 5 to 10 times more efficient than a Phantom 4 RTK, it will therefore be ideal for very precise and / or large area readings.
The second announcement of the famous drone manufacturer a LiDAR. At the start of 2020, Dji joined forces with the company Livox, specializing in LiDARs, and here is the result of this collaboration: The Zenmuse L1. This sensor integrates a mid-40 LiDAR sensor, an IMU and 2 cameras.
The LIDAR system is capable of taking 240,000 points per second, of receiving up to 3 echoes with an FOV of approximately 70 °. Designed for flights at a height of between 50 and 150 meters, you will get between 100 and 300 points per m². The system proposed by Livox has the particularity of sweeping in a “rosette” in order to optimize the recording of points.
The IMU coupled to one of the 2 cameras as well as to the drone’s RTK system makes it possible to geolocate the point cloud with great precision . Dji announces a relative precision of the results of 5cm in PPK, 10cm in RTK and an absolute precision of 10cm in PPK, 15cm in RTK.
The second camera has the characteristics of that of the Dji Phantom 4 Pro and RTK, namely 1 inch for 20 megapixels. It allows the colorization of the point cloud obtained thanks to LIDAR.
With these specs, it is legitimate to ask “why use the technique of lasergrammetry rather than photogrammetry?” “
This triple echo system makes it possible, for example, in vegetation to measure points of the treetops, branches and natural terrain. Thus, unlike photogrammetry, it is possible to know the topography of a land through a plant cover.
Another advantage: this method is based on active optical technology, LIDAR creates its own light, unlike photogrammetry which is a passive method which requires good light conditions. It is therefore possible with a LIDAR to take readings in dark environments, on smooth textures (sand, snow, etc.)
Also, the measurement rendering is almost instantaneous with a LiDAR. The Dji Terra Software will be able to post-process the point cloud with satellite observations, camera views and IMU recording within minutes. Post-processing of the photogrammetry data takes several hours.
To date, prices have not yet been communicated and the 2 nacelles will be available in early 2021. It will therefore be necessary to be patient. We can’t wait to test these 2 very promising products for many professionals: topographers, architects, archaeologists, etc …
For more information, contact us, we will be happy to exchange with you!