Traffic studies from the bird’s eye view: GAME CHANGER IN MICROSCOPIC ANALYSIS

Do you need super-accurate data about every single participant of the traffic flow? Are you looking for a non-invasive solution offering a complete telemetric data about all moving objects in the recorded image, including cyclists and pedestrians? Does your project require analysis of speed profiles and various types of acceleration (lateral or transverse), or distances between objects and even their dimensions? Are you looking for an ultimate solution in the field of microscopic traffic analytics for research of interactions, calibration of models or even development of autonomous vehicles and other scientific and commercial projects? With DataFromSky and it’s bird’s eye view using drones and AI in image processing, you get all of these functions and much more.

Why use a drone for traffic analysis?

A drone is a perfectly mobile camera with immense range of covered area, which can be “installed” within a few seconds. One drone is enough to cover the whole intersection or even a whole transport corridor. A bird’s eye view is the most suitable option even in regards to metering of dimensions, speed and acceleration of objects on the road. With this type of view, you can get data so accurate that will be able to detect change of gears of drivers. Interactions between drivers and links between individual traffic situations and objects can be observed much easier. Thanks to its accuracy, complexity, range and availability, drone-captured traffic data open up a new chapter in the field of microscopic traffic analyses.

https://vimeo.com/230806084

Bird’s eye view is mathematically ideal when it comes to localization of ground targets. In this case a camera has the highest sensitivity even with changing of object position within the image. With growing incidental angle, this sensitivity decreases rapidly. This is the reason why it is not possible to acquire highly accurate data with cameras located alongside the road. Another advantage of the bird’s eye view is with natural minimization of object alignments and occlusions, making the metering even more accurate as the objects are always clearly visible. With this type of camera view, localization accuracy of a vehicle reaches up to 15 cm, with the measurement error under 2% and measuring to 30 cm in dimensions (according to the configuration). If you want extremely accurate not only in counting of object in classes (O/D matrix, TMCs), but also in telemetric data (speed, acceleration, object size), the bird’s eye view is a perfect option. With this accurate telemetric data, it is possible to detect dangerous situations, calibrate models of driver behaviour or even count the capacity of crossroads and much more.

*A dependence of localization accuracy on the angle of incidence and flight altitude. The plot shows the area which is possible to be covered* *with the given localization accuracy of vehicles up to 20 cm, using the 4K camera. Source:* *DataFromSky*.

*Accuracy of localization with a view from a drone. Contours within the image represent the possible object localization accuracy at the given* *part of the image (up to 15 cm!). Source:* *DataFromSky*.

Up to 250 000 m² of covered area

With only a single flying camera/drone, it is possible to cover an immensely large area. From the altitude of 120 meters, you can monitor a road section of up to 250 meters in length. With 300 meters of altitude, you can cover up to 700 meters of the road, and that is a very decent area of monitored traffic area! To acquire image data even from very large traffic corridors, it only takes one device. This is why DataFromSky gives you the most detailed analysis for every single participant of the traffic flow you measure.

https://vimeo.com/231056492

Installation within a few seconds

It is definitely not easy to prepare and carry out a traffic research, especially in case of a larger intersections, as you have to install a sensor device for every single mouth of the intersection. You also have to deal with mounting, power supply and on top of this, even protect the device against robbery or vandalism. What you usually get is a series of outputs, which have to be interconnected into one coherent output, which is a very complicated and tedious process. This is not the case if you choose to use a drone for your research! All it takes to analyze the given traffic flow is to fly above the section you are interested in, and make only one video-recording, which contains everything you will need. If your project requires recordings for longer period of time, a drone can be supplied with a cable or you can swap a few drones to get the job done and give you the data you need.

Figure: *Drone on the cable — tethered solution and drone swapping . Source:* *DataFromSky*.

An ultimate solution

A video from the bird’s eye view is the cutting-edge solution also when it comes to richness of acquired data. With one take, you get microscopic data from the whole intersection. AI and smart tracking algorithms then convert the video output into an extract with super-accurate trajectories within the individual classes. With data like this, you can answer even very complex traffic research questions, such us: What are the bottlenecks of the traffic flow? Where do the conflict situations take place? What is the capacity of the intersection? How much time do the vehicles need to cross the intersection in its various parts? How many vehicles have breached the speed limit? Where do the pedestrians usually cross it?

https://vimeo.com/317429936

Goahead and try our new platform called DataFromSky AERIAL for free — the most advanced microscopic analysis tool from aerial footage!

DataFromSky is a complex system for fully automated extraction of accurate traffic data using AI and machine learning methods. It has been developed and used for the purpose of traffic studies since 2013. DataFromSky AERIAL is a state-of-the-art and fully automated service for video analysis from the bird’s eye view (drones, balloons) for acquiring of super-accurate data not only for traffic research. DataFromSky LIGHT is a service for data analysis from low-altitude cameras for standard traffic research, fully automated as well. Try and see all the traffic magic FOR FREE!

Tuesday August 27th, 2019 in News

DataFromSky — the philosophy behind

DataFromSky is a cloud-based platform for fully automated traffic analysis from videos using AI and machine learning methods. The output of the video analysis is an extract in the form of classifications of trajectories. This extract is then analyzed for various traffic tasks with numerous analytical functions it offers, fully interactively. Why is DataFromSky so unique:

● Super-accurate data suitable for traffic research (read more)

● Possibility of measuring a range of advanced parameters thanks to the trajectory-based approach

● Impeccable control of outputs thanks to extreme detail in displaying and simple export of results

DataFromSky is composed of two advanced platforms — DFS LIGHT and DFS AERIAL. The AERIAL platform serves for traffic analysis from the bird’s eye view, using the state-of-the-art technology in the microscopic view on data. LIGHT version serves for standard traffic research in mass-scale using fixed cameras and numerous functions from the AERIAL platform. Both versions of DataFromSky have their real-time versions thanks to the series called DataFromSky Embedded.

Figure: both versions of DataFromSky. Source: DataFromSky.

Created for research, developed for traffic studies

Development of the system started as a part of a research project in cooperation with BUT in 2013. The goal of the project was to use drones and advanced image processing algorithms to develop a new approach in the field of traffic analysis based on extracts of trajectories within the image. Thanks to its extreme accuracy, this project made brand new research goals, even those thought to be not feasible at that time, possible. The main aim of the project was to test the ability to localize ground objects with accuracy less than 25 cm from a drone with camera, which is suitable for advanced traffic analysis such as security audit, calculations of intersection’s capacity, research of interactions between individual participants of the traffic flow, etc.

The project entered the market under the name of DataFromSky and this allowed for further dramatic development and creation of a solution that is unique worldwide and already used in hundreds of research projects. From the very beginning, DataFromSky aims at solving the most challenging tasks in the field of traffic analysis. It is a software-based on scientific research and unique know-how, as it is constantly improving with sensitivity analysis of the system and numerous simulations as well as measurements within the real traffic. DataFromSky has by far the largest database of traffic behavior from places all around the globe, thanks to analyzing thousands of intersection nodes with extremely detailed data. Such a database background also guarantees the most advanced AI for these types of analyses, as our AI constantly improves with learning from the biggest database of traffic situations recorded with a drone, worldwide.

No black box solution

The philosophy of DataFromSky is to provide data that you can trust. This is the reason why it is based on possible highly detailed checking of the outputs of the analysis so that you can verify the accuracy of the automated analysis in a concrete video/place of yours. DataFromSky is not a black box solution, which would use only statistical data to hide the actual quality of outputs. It is actually the other way around, as with DataFromSky, you will have outputs displayed in a well-arranged and transparent manner so that you can easily evaluate the validity of data from the automated analysis. Always demand a possible control of outputs from detection algorithms for individual vehicles, as within the first five minutes of the video, you will already reveal potential system failure, which could consequently lead to an incorrect engineering decision. With DataFromSky, you have this option, as it stands on a philosophy that behind every good decision there are accurate data. No black box solution!

Guarantee of 100 % results

Collecting data is usually a time-demanding and, consequently, also rather costly process. It is not always possible to record a video in a way that an automated analysis would be able to extract all the data you need. Reasons for failure of the automated analysis can differ from bad quality of the video, unsuitable angle of recording, mechanical problem, etc. To make sure that you have positive returns on your video recording investment (which can be rather expensive in case of recording from a drone), DataFromSky offers a program called „Guarantee of 100% results.“ If you decide to use this program, our AI-based control system will come into the game and provide throughout analysis of the outputs and fix the incorrect results in a way that your traffic study is impeccable and error-free. This service can be purchased at any time in case you find out that the automated output did not match your expectations.

Data is yours — no further payments

You can download a complete analysis to your own computer and store it, therefore you can come back to it at any time in the future and without any further costs. Outputs of the analysis can be accessed, viewed and worked with offline from any place even without the internet connection. To open and view the data, use our professional application called DataFromSky Viewer, which also offers a whole range of advanced analytics including displaying of results directly in the video. With supported “views” on traffic data, you can easily set various types of configurations of traffic analyses. Thanks to this interactive application, you will also be able to immediately react to any new emerging questions related to your traffic research even during the presentation of its outputs.

Need to add a new analytic function?

DataFromSky platform is constantly under development on all levels. If there is a specific type of export or a new type of analysis that needs to be added, let us know. We have already implemented tens of advanced functions on demand and incentive of our users, while some of them are still available. We have rich experience with the implementation of request from our clients in the field of traffic analysis, as we have been involved in numerous research projects already. As a DataFromSky user, you get not only super-accurate data, but also personal approach and professional support in order to successfully realize all your projects.

Real-time analytics — DataFromSky EMBEDDED

DataFromSky is not solely about post-processing. The system also embodies devices which allow transforming literally any camera into a smart sensor. We took the AI from DataFromSky and optimized it into the smallest processing units. This is how the DataFromSky Embedded series came into being. It is a platform from DataFromSky using the state-of-the-art processors with low input power, built for the operation of deep neural networks. Do you want real-time data? Do you want an adaptive system for traffic signal control within intersections? Do you want a smart camera? Ask for the DataFromSky Embedded series.

DataFromSky has revolutionized the process of traffic data collection and analysis and unlocked new possibilities in the analysis of microscopic traffic behaviors, such as capacity evaluation and detection of risky situations — near accidents. Source: DataFromSky.

Monday August 26th, 2019 in News

How to capture a perfect drone video for your ultimate traffic survey?

The employment of drones and automatic image analysis offers new possibilities in traffic surveys, thanks to the accuracy of the measurement, the coverage range and the uniqueness of the obtained data. The camera is the most sensitive on the change of the position of ground objects in all directions when using the bird’s eye view. The dynamic occlusion between individual objects is minimalized as well.

These factual reasons together with advanced algorithms for extraction of traffic data ensure perfect information about every single participant of traffic flow.

Main aspects:

Angle of recording/ drone position
Lens speed / parasitic optical phenomena
Resolution & framerate & bitrate
Legislation

Angle of recording / drone position

In terms of accuracy of vehicle localization, the most ideal drone position is directly above the intersection — so-called TOP DOWN VIEW. At this position, the camera is the most sensitive to the change of the position of the object in the scene, therefore an extremely precise measurement of vehicle speed and acceleration is made possible. At the same time, the dynamic occlusion between individual objects is minimalized. Nevertheless, drone suspension directly above the traffic node is not always possible due to safety regulation and applicable legislation. In a situation like this, a simple rule applies — place the drone in such position that the view of the analyzed section is near to the bird’s eye view.

The procedure is as follows:

Fly with the drone to the nearest possible place in the middle of the intersection.
Ascend to the maximum possible height. However, do not cross the height where a pixel resolution of monitored objects would be less than 20×20 pixels.

There is a notable drop of the accuracy of the vehicle localization (please note that this does NOT hold for DETECTION or for CLASSIFICATION) if the angle of the view is higher than 55 degree.

A simple rule is as follows:

Angle of incidence ˂ 55 degrees = the drone should always be higher than the distance between the projection of the drone position on the ground and the center of the intersection.
45 degrees means, that the drone altitude is the same as its distance from the center of the intersection.
0 degrees means that the drone is directly above the intersection.

Figure: three situations -0° = ideal position, 45° = ok, altitude is the same as distance from the center of the intersection, 55° = drone should be higher. Source: DataFromSky.

The drone should always be positioned in that way so the dynamic and static occlusions are avoided. Occlusions increase the probability of failure of object tracking and of disengagement of continuous trajectories, in other words, one object will have a higher number of shorter trajectories. The systems are robust against short occlusions, but long occlusions can cause the disengagement of continuous trajectories into more parts.

Figure: two situations — long occlusions are problematic, but system is robust against short occlusions. Source: DataFromSky.

Try to keep the drone in the same position when recording. Do not change the drone altitude nor the camera take and zoom when recording. All parameters should have been set before the start of recording. For high-quality recordings, it is better to use a camera with ultra-wide range lens, not with fish eye lens, which causes high distortion especially noticeable at image margins.

Figure: two situations — try to avoid changing drone position when recording, and also avoid recording against the strong source of light. Source: DataFromSky.

Parasitic optical phenomena / lens speed and scene acuity

When recording, try to avoid parasitic optical phenomena in camera take, which may be caused for instance by the sun or another strong source of light. Always turn the camera in such a way that this strong source of light is not recorded by a camera. Takes of scenes with sky should be avoided.

Figure: three situations — nok: too bright (left), ok: ideal (center), nok: blurred due to low exposure time (right). Source: DataFromSky.

Focusing should always be carried out in such a way so the vehicles / objects are not blurred in a take. At the same try to set the camera in such a way so the objects are contrasting but not over-bright nor too much dark. This should be considered even at the expense of other parts of the shot. In case of night recordings or the recordings at the low light conditions, it is important that besides vehicles also the traffic elements are visibly identifiable.

Figure: nok — four situations. Source: DataFromSky

An image from camera ought not to be deformed by shaking (caused by drone motor vibrations) as a majority of cameras use the rolling shutter function to load image data in rows. If the vibrations are not sufficiently damped, their effect will occur in the form of shaking in between rows within the image. This shaking then lowers the accuracy of the used stabilization methods. Motor vibrations can be damped with suitable damping elements.

Resolution, FPS and bitrate

Resolution is a key attribute for the detection and classification of objects. The minimal size of an object within the image (for its detection) is 15×15 pixels, while for its correct classification, a resolution of 30×30 pixels is needed. A recommended size for objects within the image is between 30×30 pixels up to 150×150 pixels in the whole monitored area within the intersection. Object size depends on the following parameters:

Image resolution
Drone’s position — altitude / angle
Mutual position of a monitored object and a drone/camera

A common rule is that if you record the video with maximal resolution used by modern cameras these days (e.g. 4K or 5K), it will be sufficient, as it is always possible to reduce the quality, but not the other way around. The following is a simple guide for resolutions:

A distance with oblique angle < or = 120 meters -> FULL HD (1920×1200) resolution will be sufficient, but it is better to use a higher resolution (2k/4k)
A distance with oblique angle > 120 to 300 meters -> use 4K video resolution

Figure: min resolution. Source: DataFromSky.

Framerate represents the sampling frequency of measurements. For vehicle speed up to 70 km/h, it is possible to use a framerate 15FPS. For vehicle speed over 50 km/h, it is better to use sampling frequency around 25 FPS. Sampling frequency over 35 FPS has no further positive effect on the quality of measurement, but only makes the video and its size unnecessarily large. The final quality of the video is also greatly influenced by bitrate settings. At FULL HD videos, coded with the H264 coding format, the bitrate should be at least 10 Mbit/s. At 4K videos, coded with H264, a minimum of 20 Mbit/s is necessary. The video must not be blurry, pixelated, or any signs of damage caused by compressing algorithms.

Tip for beginners: record video with 4K resolution and 25 FPS with standard bitrate (approx. 20 Mbit/s)

By following these rules you will get suitable videos for automatic processing to extract super-accurate data for your traffic study. Before the final recording, we recommend to carry out a test flight or make a few sample recordings in order to get acquainted with the basic principles. If you have any questions, do not hesitate to contact us.

Legislation & safety

Before the recording itself, always check what are the legal conditions for the recording of traffic in your country. Legislation differs from country to country and sometimes it is required to use services of a professional pilot with necessary permissions. There is also a predefined maximal flight altitude for drones in many countries (usually 120 meters), as well as a minimal safety distance (usually 50–120 meters away from objects). If the recording is carried out by a professional pilot, it is possible to gain permission even at different conditions. Always take the possible risks of technical failure into consideration.

If you do not want to carry out the recording by yourself, do not hesitate to contact us. We have a network of experienced and certified pilots all around the globe within the DataFromSky PILOT program. We will provide you with consultations about your project and its requirements in order to find the most suitable solution. We can also carry out the recording for you if necessary.

Monday August 19th, 2019 in News

DataFromSky Aerial analysis in the Netherlands

ROELOFS, an exclusive partner of DataFromSky in the Netherlands, carried out comprehensive traffic research of a traffic node in the province of Gelderland. The data collected with a drone and analyzed by DataFromSky Aerial platform will be used as the main source for a study for planned major maintenance in 2021.

The drone was deployed at the roundabout of the N318 Misterweg in Winterswijk. When preparing for the major maintenance of this road, it is also being examined whether road safety can be improved. Besides other things, DataFromSky offers many great features for analysis of traffic behavior and road safety, including analysis of risky situations or near collisions, calculation of Time to Collision, Heavy Breaking and analysis of many other factors.
For many years, the data were collected by people standing on the side and observing the traffic, then the data were evaluated manually. This process was lengthy, inefficient, and had low-precision data output. All of these shortcomings are overtaken with a new modern approach combining 2 great technologies of today world: drones and AI.

Traffic flow – trajectories: green – cars, yellow – heavy vehicles, red – bicycles, blue – pedestrians

Read a post in “de Gerlander” online magazine.

Thursday April 18th, 2019 in News

DataFromSky supported by Technological Agency of CR

The whole team of DataFromsky is happy to announce that the Technological Agency of Czech Republic has supported our project in the international programme “Delta”. We have succeeded with a project called “Traffic management platform development for big social events using a drone based on visual sensing techniques”.

The aim of the project is to develop an aerial artificial intelligence – a comprehensive mobile platform for real-time monitoring of traffic and big social events, using a UAV and advanced methods of computer vision based on artificial intelligence. The goal of is to develop a fast-deployable system for real-time surveillance of the defined area with automatic object recognition and classification, monitoring and automatic motion evaluation. Extracted image data will be analyzed in order to detect some nonstandard behavior, congestion, zone disturbance, and other user-defined parameters. Traffic data will be used for the immediate evaluation of traffic using simulations. The system has a high potential of being deployed in the traffic monitoring segment (police drone) or in a security field.

The novelty of the solution is in the implementation of intelligent UAV, which will provide information about moving objects during the flight operation, and interpret this data in the context of a particular application. This highly mobile platform is used wherever it is necessary to analyze the real-time situation.
The project is a multidisciplinary task that combines the latest trends in mobile robotics (UAV), computer vision, and automated telemetry analysis. The system will be integrated into a higher unit to evaluate traffic situations and traffic plans based on simulations.

Friday March 29th, 2019 in News

Advanced roundabout analysis with DataFromSky

Need to perform operational analysis of a roundabout based on the real traffic data? DataFromSky supports the calculation of a wide range of operational parameters such as critical-gap time / follow-up time / critical headway etc. Get super accurate traffic data for your research and/or traffic study regardless of the type of intersection using DataFromSky.

Monitoring and analyzing traffic behavior at intersections and roundabouts is an important task in the present traffic research. Trajectory analysis provides higher-level situational awareness by understanding the behavior of every traffic element in the analyzed scene due to the rich information contained in its own motion. DataFromSky is able to extract trajectories and their spatial-temporal statistical characteristics automatically with very high accuracy and high sampling frequency. Accurate traffic data is essential in any modern research in the field of traffic analysis such as:

Motion pattern classification, Abnormal motion pattern detection,

Traffic mining, Prediction of non-recurrent short-term traffic patterns,

Developing next-generation driving behavior models, Trajectory clustering,

New perspective on merging behavior – gap seeking instead of acceptance behavior,

Evaluation of traffic simulation methods / models on real data, Capacity estimations,

Traffic congestion studies, Travel-time, Vehicle counting, Trajectory based analysis,

Road safety, Pedestrian and cyclist safety, Road-user interactions, Smart analysis,

Driver decision making and complex multi-actor interactions, Traffic instabilities,

Empirical evidence of multi-anticipation (drivers have multiple leaders),

Study of behavior adaptation effect.

Friday February 15th, 2019 in News

Automatic traffic analysis on an area of 55,000 square meters? Possible!

Object counting and classification, gathering speed & acceleration data, O/D matrix and much more on an area of 55,000 square meters. DataFromSky is the name of the most advanced Aerial AI for an understanding of movements in videos taken by drones, balloons or helicopters.

Take a look at another nice shot from Italy taken from a balloon. The length of the video is 2 hours and 31 minutes. DataFromSky tracked 5 455 objects (4881 cars, 325 medium vehicles, 77 heavy vehicles, 68 buses, 101 motorcycles, 1 bicycle, and 2 pedestrians) with the average speed of 12.24 km/h and the total distance traveled by all participants reaching 1 847 km! Do you need extremely accurate data about traffic flow? Contact us!

Thursday February 7th, 2019 in News