Published by Daniel Stofan on
Almost every company from transportation industry tackles the challenge of efficient traffic data collection. Still, just a few keys know how’s will maximise the benefits for the good of their projects and all parties involved.
%20(1).webp?width=679&height=382&name=1_KNF2XcFCf2XoX_8OJ8m0oQ%20(1)%20(1).webp)
GoodVision is a traditional provider of comprehensive traffic analysis platform combining automated data retrieval from video footage with advanced data analytical tools for modellers. The platform has an excellent advancement to how traffic surveys are traditionally achieved. Today it is evident that automated methods of collecting traffic data using artificial intelligence are in the spotlight of transport planners and municipalities. What does this mean for the current processes, how does it look when you decide to implement such technology to your traffic projects? This article describes it in 3 simple steps.
STEP 1: Retrieving real-world traffic data automatically from cameras
A video or a camera feed is needed to capture the traffic in the location. Camera is a very convenient sensor, very underestimated in the past, but can be greatly utilized using the latest technology. Typically traffic surveyors are placing the temporary or permanent cameras to capture the real-world situation for analysis. Although the on-camera requirements positions for video analytics are more strict than for human traffic counters, it brings enormous advantages when done correctly.
A professional team of GoodVision consultants supports you when choosing the camera position, performs the verification and is walking you through the process. Live camera feed or recorded video files are then conveniently uploaded by surveyors to GoodVision’s platform for data extraction. GoodVision’s method of traffic data retrieval uses on proprietary Artificial Intelligence recognising individual traffic attendants in the video footage by its visual appearance. The whole process is automated, with no humans involved. What are the benefits?
- Independence of the geographical location of the survey
- No need for adaptation of the system before the survey
- Stable data accuracy as the human factor is removed
Picking the source of traffic video
In the first iteration of the service, GoodVision was relying on video recordings and its direct upload to the platform. Now the platform is getting significantly smarter. GoodVision Video Insights allows to provide video content for processing in the following ways:
- Video recordings
- Real-time camera feeds — connect a video stream from the camera directly to Video Insights, schedule the traffic survey according to your needs, and get data updates on the fly. Read more in our article.
- Drone video footage (flying up to 250 meters high)
- Time-lapse footage (e.g. from Brinno cameras)
GoodVision’s approach of traffic capturing has multiple benefits over traditional methods:
- Hardware independence — system can work with video input from virtually any camera manufacturer.
- Speed of data turnaround —traffic data from video footage is obtained within unbeatable 1 HOUR, regardless of the video duration.
- Multipurpose data —you’re getting not just traffic counts, but full versatile data allowing for repetitive analysis allowing to use the data for other purposes in the future, like traffic modelling.
- Data accuracy — is 95% throughout various weather conditions when the requirements are met.
Once the data from the cameras are extracted, it is ready for analysis in various ways. Often, surveyors can perform traffic volume counting and reporting in the platform by themselves, leaving only the advanced analyses to modellers. GoodVision’s customer service will make sure you have achieved the analyses correctly to get the best data accuracy possible. GoodVision’s users experience stress-less delivery of traffic jobs within deadlines. The automation element in the platform covers the majority user’s manual tasks and reliably delivers the extracted data within 1 hour upon upload.
STEP 2: Sharing digital data to traffic modellers
The most common scenario today is traffic surveyors sharing collected video footage with traffic volume spreadsheets to traffic modellers for review and manual analysis. This process is becoming smarter with GoodVision — surveyors share all extracted digital data on a click of the button within the platform directly to modellers . This is a win-win collaborative solution for both parties.
Modelers benefit from digital traffic data in the most significant detail ever for calibration of their traffic models, while significantly cutting costs on manual video review. And surveyors benefit from automation of traffic counting and increased productivity of their staff.
Traffic data in this digital format allow for a wide variety of further analyses. This inconspicuous collaboration feature makes the project delivery as simple as sending the email.
STEP 3: Analysis of traffic data for model calibration
GoodVision Video Insights is the first platform of its kind, allowing such interactive deep traffic data analytics while saving time on manual video review and calculation of calibration parameters. It is a platform where surveyors and modellers can collaborate on their projects.
While peak period analysis is still the most important for traffic modelling purposes, modellers are allowed to perform the analyses over longer samples of the data as their manual review is not a limiting factor anymore. Just take free-flow analysis into account — it is often performed off-peaks. Video Insights retrieves and stores all physical aspects of the traffic attendant’s movement in the form of trajectories. This nature of the data allows for getting detailed traffic parameters leading to the calculation of crucial calibration parameters. As an example, modellers can:
- Use headway times and volumes to calculate saturation flows automatically in the app
- Use traffic delays and gap acceptance data to analyse the service level on intersections
- Utilise lists of individual vehicle passages together with signal timing intervals to identify red-light violations
Modellers can share the same data within their team and perform analyses separately. Even it this phase, GoodVision’s consultants are ready to assist with the platform features including the advanced scenarios specification.
The versatility of the retrieved data allows for repetitive analyses of various types over the same set of data. The results are delivered to the user online. Can you imagine how many staff-hours of your traffic analysts are saved on removing the tedious manual tasks from them?
Visual representation of the traffic data is another key factor of GoodVision Video Insights allowing visual interactivity of the analyses, providing exact vehicle and pedestrian traces, heat-maps and time occupancy maps. This is extremely valuable to visually identify the anomalies in the traffic flows, jay-walkers or illegal manoeuvres.
Visual representation of the traffic data is another crucial factor of GoodVision Video Insights allowing visual interactivity of the analyses, providing exact vehicle and pedestrian traces, heat-maps and time occupancy maps. Identification of anomalies in the traffic flow is extremely valuable visually as well as detecting jay-walkers or illegal manoeuvres. Following are the critical traffic behavioural parameters which can be collected from the roadways by GoodVision:
- Multi-modal turning movement counts
- Vehicle classification into 8 classes
- Pedestrians and bikes
- Peak-hours and free-flow intervals
- Visualization of exact object traces
- Travel time and speed
- Occupancy times and delays
- Headway time
- Jaywalkers, desire lines, illegal manoeuvres
- Red-light runners
Reporting the data into traffic simulation systems
Almost everything in Video Insights can be exported, including graphs and visual maps. It can provide various types of built-in tabular data reports including standard traffic movement counts, Origin-Destination matrices and vehicle lists with timestamps and detailed behavioural parameters of each vehicle in the flow.
GoodVision Video Insights is designed to work closely with traffic simulation systems like VISSIM or LINSIG. It provides data reports allowing its direct input to these systems. Anytime, i.e. when configuration errors made, a modeller changes the filters and re-generates the data reports immediately without the need of video re-processing.
Conclusion
Don’t waste your precious time with manual traffic counting. It’s not just about the data turnaround time. Just imagine, with AI-powered software platform like GoodVision’s you can get way more than Excel spreadsheets. You are getting the complete raw data about traffic behaviour which can be utilised for model calibration directly. With data privacy as another essential factor, GoodVision Video Insights is GDPR compliant. It uses machines to process the video footage in a protected cloud environment with no third parties having access to it and videos deleted immediately after processing.
