The solution from Bluecity features Lidar technology and AI algorithms that capture real-time multi-modal traffic data.

The solution is non-intrusive and differentiates itself by providing reliable data in any weather or lighting condition. What’s more, unlike cameras, the system collects anonymous data, avoiding any issues regarding privacy.

To encourage the adoption of Bluecity’s solution, the company has been implementing ongoing studies to confirm the accuracy of lidar technology. In the first part of the study an intersection that had been equipped with a 3D lidar sensor, in Kelowna, British Columbia, was monitored in 15-minute aggregation intervals over the course of three hours. The ground truth (data collected by an analyst) was compared to the 3D lidar sensor data processed by the AI algorithm to see if the algorithm mirrored the ground truth.

The intersection had been identified by four quadrants: North, South, East and West, and cars had been counted as they moved from one quadrant into another. The cars could move in a total of twelve different trajectories, for example, north to south, north to east, south to east etc.

Primary Results

Over the three hours, the accuracy of the algorithm ranged from 90.57% to 99.21%, with an average of 95.84%, but it was noted that two trajectories (north to west and south to east) showed accuracy that was lower than the other trajectories.

The Bluecity team determined that in the case of the north to west trajectory, the height at which the 3D lidar sensor was positioned was responsible for undercounting the cars. In the case of the south to east trajectory, they determined that an overcounting of cars resulted from noisy data, or data that the system was interpreting incorrectly. The team changed the size and position of the virtual loops for that trajectory and in both cases, retrained the algorithm by feeding it more data.

The next step was to repeat the study and see whether the 96% average accuracy could be improved. The team performed a quick 15 minute data collection and found a remarkable improvement – the accuracy increased to 97.50% and in each of the two problem trajectories, the new process yielded a delta of only 1-car.

New Data

New Accuracy

Knowing that a 15-minute follow-up may not be enough to validate improved accuracy, the Bluecity team undertook another study whereby they collected data for 6 hours using the same methodology, i.e. aggregating data in 15-minute intervals and comparing ground truth with the data from the algorithm for the two trajectories in question.

This time, the north to west trajectory only undercounted 3 cars out of 60 for 95% and all 204 cars were counted in the south to east trajectory resulting in 100% accuracy. The total accuracy for the two trajectories was 97.5%



Additionally, the Bluecity team analyzed a second intersection to see if the accuracy remained as high as it did at the Kelowna intersection. The second intersection was in Cote St. Luc, Quebec. Due to high volumes of traffic during rush hours, the city of Cote St. Luc was studying how to improve traffic congestion at the intersection. To help with planning, the city installed a 3D lidar sensor to collect data on road usage. At this particular intersection, there was no traffic arriving from the west so there were nine trajectories studied in total.

Data was monitored in 5-minute aggregation intervals over the course of two hours. When the ground truth was compared to the lidar data, the Bluecity team found that the average accuracy for this intersection was 97.9%

It was now clear that the accuracy of 3D lidar sensors had been confirmed. Cities, as well as engineering and consulting firms, could all adopt Bluecity’s new lidar technology with confidence.

Bluecity also plans for an ongoing series of studies that will investigate variables such as speed, classifications, weather and lighting conditions, and safety analytics. The objective of these studies is to confirm the accuracy, reliability, and broader insights available from lidar sensors.

Bluecity’s goal is to make streets safer for pedestrians, cyclists and motorists. We envision a world where traffic accidents are rare, driving to work is a pleasure and carbon emissions are reduced. 

Download now