Testing Seyond's Lidar Solution Under Winter Conditions 

Seyond and its partner Normivalaistus have recently conducted a validation test in Finland to assess the performance of Seyond’s ITS Management solution called SIMPL-Intersection—comprising two Falcon K Lidar sensors and Seyond’s advanced perception software—in detecting and classifying different types of vehicles under challenging winter weather conditions.  

The test focused on the accurate detection of cars, buses, and trucks passing through two virtual detection zones defined on Seyond software. 

Test Conditions

The validation test was conducted in a winter environment featuring: 

  • Overcast skies 
  • Snow 
  • Sleet 
  • Ice 
  • Wind 

These conditions were chosen to simulate the typical challenges faced in urban mobility scenarios during winter months in Nordic countries. 

Methodology

Virtual Detection Zones: Two virtual detection zones were configured at the stop bar using Seyond perception software. 
Ground Truth Validation: Six 15-minute recordings were used as part of this test. A manual validation process served as the ground truth for comparing Lidar detection results. 
Vehicles Detected: The test monitored and validated the detection and classification of cars, buses, and trucks. 

Results

The performance of the Lidar-based ITS solution was evaluated based on the deviation between detected and manually validated vehicle counts. 

Accuracy report data - winter conditionss

 

Accuracy Rate: 97.7% 

Misclassification Observations: 

  • Three big vans were classified as cars. 
  • A tractor was classified as a truck. 
  • Some vehicles coming to the intersection at the very end of the clip may have not made it to the virtual loop which could explain some of the deviation. 

The slight deviation of -2.27% demonstrates that Seyond’s Lidar solution maintains high accuracy in detecting vehicles despite challenging winter conditions. The observed misclassifications (ambulance as a car, tractor as a truck) indicate areas for refinement in object classification algorithms to enhance performance further. 

In conclusion, the validation test under winter conditions confirms the reliability, accuracy and cost-effectiveness of Seyond’s Lidar solution for advanced traffic management applications. The system’s performance underscores its potential to provide accurate detection for intersection actuation, making it a valuable asset for cities of all sizes striving for improved urban mobility. 

Accuracy Report

Accuracy Test: SIMPL has a 99% accuracy rate in vehicle counting at the stop bar in Florida

READ MORE
CASE STUDY

SIMPL uncovered critical insights into pedestrian safety at mid-bock crossings.

READ MORE
Comparison Article

Comparing Inductive Loops and LiDAR Technology

Visit

Ready to see it in action?