#cleanworld challenge launched

June 25, 2019

#cleanworld challenge launched by Novarum Sky and artificial intelligence PowerBrain.Shop®

Next-generation Artificial Intelligence software business PowerBrain.Shop® and Drone solution provider Novarum Sky announced launch of #cleanworld challenged.

Eindhoven, June 25, 2019 /PowerBrainShop/ -- PowerBrain.Shop®, São Paulo, June 25, 2019 /Novarum Sky/ The Drone solution provider Novarum Sky and their artificial intelligence (AI) technology partner PowerBrain.Shop® announced they joined forces to help driving the development and application of environmental protection solutions based on latest AI and Drone technologies.

Inspired by the passion of the Fridays for Future (#fridaysforfuture) movement and the Trashtag challenge (#trashtag) the Brasilian CEO of Novarum Sky, Mr João Marcelo Correa, and the German CEO and Chairman of the Board of PowerBrain.Shop®, Mr Samuel Thomas Staehle, invite the public to help developing the next generation of artificially intelligent cleaning technology solutions to help protect our environment.

Today publicly announced at the DRONE SHOW at the Frei Caneca Convention Center in São Paulo, Brazil, the #cleanworld challenge tackles the pollution of our planet including oceans and land, forests, rivers, parks and deserts. Everyone is invited to help recording training video data with smart phones, cameras, drones and even robots showing trash in the regional environment. Afterwards it should be tagged with the hashtag #cleanworld and uploaded to public video sharing services.

This way artificial intelligence development teams around the globe can start building AI for autonomous drones and robots that help detect and remove trash around the world. Such videos also enable various additional technologies to be developed such as environmental pollution monitoring and alarming drones, environment cleaning robots and possibly even underwater robotics solutions.

The #cleanworld challenge also demonstrates to political and public decision makers that there is infrastructure technology available today which can help keeping cities, parks, forests, lakes and rivers clean on the long term – even at hard-to-reach places and powered by pollution free clean energy.

"As environmental protectionist ourselves we want to make our world a better place by embracing latest artificially intelligent technologies to help building environmental cleaning solutions" says Samuel T. Staehle, founder and CEO of PowerBrain.Shop®.

"Novarum Sky's latest Drone solutions significantly benefit from using the capabilities added by the portable AI PowerBrains™ of PowerBrain.Shop® for in-flight trash detection in real time video data. From our perspective the combination of our technologies is a major milestone for the automation of environment protection." says João Marcelo Correa, founder of Novarum Sky. 

About POWERBRAINSHOP Holding Corporations

PowerBrain.Shop® is a fast-growing software vendor with a European team and a global footprint, developing the next generation product development and training platform for Artificial Intelligence and powerful AI software 'brains' – its AI PowerBrains™. It pushes through the impossible by enabling the next disruptive step of the mass market adoption of AI by advancing easy AI development, training, validation and quality assurance, deployment and monitoring at various use cases scenarios across different industry verticals, businesses and organisations. As thought leader in Artificial Intelligence, we partner with those who move the world forward.

For more information, follow us at www.PowerBrain.Shop , on LinkedIn, on YouTube and on Twitter (@ShopPowerbrain).

About Novarum Sky

Novarum Sky is a drone technology solution and services provider with an international team of engineers and software professionals developing industrial solutions for aerial vehicles (UAV). It offers innovative, efficient and safe ways to carry out aerial inspections, direct georeferencing and analytics.

For more information, follow us at www.novarumsky.com and on Twitter (@novarumsky)

<Video about the project here>

Is GNSS receiver L1/L2 required on the UAV for RTK-PPK techniques?

March 6, 2019

The use of UAV to assist in the Topography activity grows every year, and the aerial photogrammetry technique is widely used to make geo-referenced maps with centimeter accuracy. But do you know the difference between accuracy and precision in the science of cartography? There are concepts that are best understood with image. You will understand the difference between precision and accuracy observed in this figure:

 

Figure 1 - Accuracy and Precision *

Direct Georeferencing (DG) using photogrammetry software with georeferenced photos via GNSS receiver of low precision and low accuracy (error around 3 meters) and photos superimposed in 80%, it is possible to make a georeferenced map with centimeter precision, but not it will be possible to make a georeferenced map with centimeter accuracy. The map will be offset with respect to the actual position on the planet and this error will be proportional to the error of the GNSS measurements used to geo-reference the photos, ie if the GNSS RMS error in the horizontal coordinate is 2 meters and in vertical of 4 meters the map will be with the same error. Precision will be best the lower the Ground Sample Distance (GSD) and the greater amount of pixels superimposed for map processing.

In the DG technique it is necessary to estimate the position of the UAV at the time of acquisition of the photo independent of the GNSS Rover receiver frequency and the flight speed. Current DG systems use PPK or RTK techniques and perform algorithms for more accurate position calculation through data generated by the IMU - Inertial Navigation System. In general, the IMU normally operates at the frequency of 100 Hz, so it provides data every 10 ms to calculate the estimated Drone position. These GD systems also require additional software, proprietary boards that integrate the GNSS Rover L1/L2 receiver into the IMU, and the cost can exceed the cost of the UAV solution.

To obtain a georeferenced map with high accuracy via Direct Georeferencing (DG) PPK, a high level of accuracy and precision is always required to correct the position of the camera/sensor at the moment of the photo exposure, so that the RMS error is lower at 10 cm. The PPK-RTK techniques is not enough. If you want to know more about this subject see: UAV and Direct Georeferencing (DG) solution x RTK. 

Experiments using L1 and L1/L2 GNSS receivers for Rover RTK

Now the question is, do I need GNSS on the UAV that operates on the L1 and L2 frequencies to increase the accuracy and precision of the position of each photo? Based on the result of technical experiments, in my opinion the answer is no. The L1 / L2 GNSS receiver is much more expensive than the L1 GNSS receiver and does not offer any advantages that justify the investment for the GNSS Rover RTK receiver. With new software it is possible to perform similarly on both types of receivers. Rather than explaining this by theory, which is relatively complex and massive, it is showing the practical results that compare the two technologies. The figures below show the results of the position of the different GNSS Rover RTK receivers, using the same antenna, installed on the car that runs through a city with trees and buildings. When the RTK / PPK solution exhibits centimeter accuracy, it means that it is in FIX mode:

 

Figure 2 - u-blox NEO M8T L1 frequency with RTK via open source software RTKLIB. FIX Rate= 94,4% of the time **

 

 

Figure 3 - GNSS Receiver ComNav using the L1 and L2 frequencies with internal RTK software, FIX Rate = 94.9% of the time **

 

Figure 4 - GNSS Receiver ComNav using only the L1 frequency with RTK software through RTKLIB, FIX Rate = 98.9% of the time **

In the UAV application, the GNSS Rover receiver antenna is above the trees and buildings and therefore has better reception of the satellites, so the total time with the RTK FIX system would in this case be very close to 100%.

Using appropriate RTK / PPK software, the technical differences between the L1/L2 GNSS rover receiver with a GNSS L1 u-blox Rover receiver are very small. The advantage of the L1/L2 GNSS Rover Receiver is that it reaches the first FIX faster than the GNSS L1 receiver, however, there are RTK algorithms that allow you to reach the FIX in less than 5 minutes with the UAV on the ground before taking off, this setup time is most often insignificant in UAV mapping applications.

Does using the GNSS Base receiver only with L1 frequency have any advantages for RTK system?

Another question you should be asking. The GNSS base receiver for RTK/PPK solution which is the GNSS receiver reference for precision and position accuracy of the GNSS Rover receiver, can be of the L1 only type? The answer is yes, but the application is very restricted. The GNSS L1 Base receiver achieves centimeter accuracy only after 20 hours of continuously receiving satellite data and executing the Precise Point Positioning (PPP) technique, and only then will GNSS Base L1 be ready to be used in conjunction with the Receiver GNSS Rover RTK. The GNSS Base L1 / L2 receiver needs only 4 hours on performing the PPP technique to achieve centimeter precision and accuracy. So I recommend using the base GNSS receiver that uses at least the L1 and L2 frequencies to use in conjunction with the UAV GNSS Rover RTK L1 receiver.

Is RTK / PPK solution sufficient to make georeferenced maps with centimeter accuracy?

Using RTK / PPK solution with receiver that guarantees FIX for more than 90% of flight time, with horizontal and vertical RMS error less than 10 cm, it is guaranteed that it can do Direct Georeference (DG) and obtain georeferenced maps with centimeter accuracy ? The answer is no. That's not enough. If you want to know the answer on that subject. See my other article the solution: UAV and Direct Georeferencing (DG) solution x RTK.

 

 Download the overview brochure of DG RTK-PPK

João Marcelo Corrêa                      Juliano Grigulo

CEO at Novarum Sky                     CTO at Novarum Sky

 

* Figure from article: Case Study: Accuracy vs Precision?

** Graphs from the article: Post-processing ComNav receiver data with RTKLIB – a more in-depth look

 

 

New High Precision Direct Georeferencing (DG) solution using UAV x RTK

February 22, 2019

The startup Novarum Sky is launching a new solution for Direct Georeferencing (DG) for use in fixed or rotate wing drones to enable georeferenced maps with very high accuracy and precision, better than 5 cm.

But before explaining about the DG Novarum Sky solution, do you know the difference between DG and RTK or PPK ? They are different techniques. In summary, it can be said that DG uses RTK or PPK technology to determine Drone's position at the moment of recording the photo, but this is not enough to make maps with precision and centimeter accuracy without the use of ground control points (GCPs). In PPK or RTK, UAV position calculation is done periodically and depends on the satellite data acquisition frequency on the Rover GNSS receiver installed on the UAV, so a 20 Hz GNSS receiver calculates the position of the UAV with centimeter accuracy at every 50 ms. The time of 50 ms seems little but a Drone flying at 10 m/s in the horizontal direction travels 25 cm in 25 ms, so in this case the average error of the Drone's position at the moment of the photo will be 25 cm, so the accuracy of the map will be decimetric. In order to obtain a Georeferenced map with centimeter accuracy using only the RTK or PPK techniques, it will be necessary to use GCPs for later use in the photogrammetry software that corrects the coordinates of the photos to the centimeter level. The lower the frequency of the GNSS Rover receiver and / or UAV high flight speed, the worse the accuracy of the RTK or PPK solution. The DG solution normally saves 50% of the on-site work time and 50% of the map processing time, because GCPs are not necessary and the lateral overlap of the photos can be only 40% and longitudinal 60%.

In the GD technique it is necessary to estimate the position of the UAV at the time of acquisition of the photo independent of the GNSS Rover receiver frequency and the flight speed. Current GD systems use PPK or RTK techniques and perform algorithms for more accurate position calculation through data generated by the IMU - Inertial Navigation System. In general, the IMU operates at the frequency of 100 Hz, so it provides data every 10 ms to calculate the estimated Drone position. These GD systems also require additional software, proprietary boards that integrate the GNSS Rover L1/L2 receiver into the IMU, and the cost can exceed the cost of the UAV solution.

After three years of research and development, Novarum Sky has developed an innovative DG solution that does not use proprietary hardware, but running RTK or PPK solution on Raspberry Pi, L1@5Hz GNSS receiver configured to receive Raw data from three constellations and new software that corrects the position of the Drone at the time of exposure of the photo with sampling frequency of 2000 Hz. Novarum Sky DG does not use IMU, however it provides the coordinates of the photos with accuracy and precision better than 5 cm at the end of the workflow without the use of GCPs. Another important factor is the time to FIX RTK that in the DG Novarum Sky solution is only 3 minutes. The base GNSS receiver can be accessed via the Internet via the NTRIP protocol or locally for the RTK or PPK solution. 

If you want to know if the manufacturer's DG solution works to make georeferenced map with centimeter accuracy, it must provide a georeferenced map report made by first-line photogrammetry software. In the report you must inform at least: 1) that GCP was used only to verify the accuracy of the position of the photos and 2) the RMS Error X [m], RMS Error Y [m] and RMS Error Z [m], all RMS Errors must be under 10 centimeters.

For example, Novarum Sky made a geo-referenced map using EtheCopter DG system on a Mikrokopter Hexa Drone flying at 8 m/s, Sony HX60V camera, three GCPs for verification, and Pix4d's photogrammetry software. The results were a georeferenced map with GSD of 1.05 cm and RMS error at coordinates X = 3.20, Y = 3.26 and Z of 3,47 cm. See the report prints:

 

 

During the months of March and April 2019, Novarum Sky will sell twenty units of the DG at a promotional price of US$ 1,500.00 plus shipping cost to companies to evaluate that are interested in providing the solution to the market in O&M or resale. Included documentation and all necessary support. We will be very happy to attend. Please forward the request to the email: This email address is being protected from spambots. You need JavaScript enabled to view it.

 Download the overview brochure

João Marcelo Corrêa                      Juliano Grigulo

CEO at Novarum Sky                     CTO at Novarum Sky

 

 

 

Our Products

The EtherCopter Direct Georeferencing System

The EtherCopter Direct Georeferencing (DG) System is the second of two currently available products by NOVARUM SKY. Used for the execution of Georeferenced Map applications it does so with high accuracy better than 5 cm through the Direct Georeferencing technique, being a low cost solution when compared to equivalent solutions on the market. Novarum Sky DG solution running RTK or PPK solution on Raspberry Pi, L1@5Hz GNSS receiver configured to receive Raw data from three constellations and new software that corrects the position of the Drone at the time of exposure of the photo with sampling frequency of 2000 Hz. Novarum Sky GD does not use IMU, however it provides the coordinates of the photos with accuracy and precision better than 5 cm at the end of the workflow without the use of ground control points (GCPs). Another important factor is the time to FIX RTK that in the GD Novarum Sky solution is only 3 minutes. The base GNSS receiver can be accessed via the Internet via the NTRIP protocol or locally for the RTK or PPK solution.

the

 

The EtherCopter Inspection

The EtherCopter Inspection is one of NOVARUM SKY's two products, which are currently available. IoT and Remote USB IP are integrated in order to increase efficiency of detailed industrial inspection, allowing a 2.5 times faster inspection report. In addition it enables the inspector to operate the camera or any other sensor that is installed in the Drone as well as to communicate with the Drone pilot through Full HD video and voice, using only the Internet and Web Browser interface from anywhere in the World.

ether