Recognition Solutions for Traffic Flow Counting & Vehicle Types

A cost-effective LiDAR, the TF03 features a high frame rate, good ambient light immunity, small volume, high accuracy, and a detection distance of up to 180 m. It has a specifically designed algorithm for the counting of traffic flow and the recognition of vehicle types, analyzed from the performance index.

The software can realize vehicle type recognition and traffic flow counting functions. As there is hopping in the LiDAR’s Dist, counting is triggered when a vehicle passes. The counting of traffic flow is then finished by this data. Using the neural network algorithm together with the LiDAR, the recognition of vehicle types will be completed by the data.

1. Traffic Flow Counting

When the LiDAR is facing down towards the ground, LiDAR dist is displayed. Hopping is displayed when vehicles pass through the measurement range. This triggers the traffic flow counter, enabling counting to be finished.

2. Vehicle Type Recognition

Basic parameters of a vehicle, such as its outline, length, and height, can be calculated from the changing rule of the Dist from LiDAR. The recognition of vehicle types can be realized via the related machine learning algorithm.

The combination of natural network technology and the decision-making tree enables the realization of the vehicle type recognition algorithm. The seven types below can be classified:

  • Hatchback
  • MPV
  • Minibus
  • Sedan
  • SUV
  • Bus
  • Truck

System Creation Solution

The figure below shows the solution:

Two LiDARs ought to be pointed towards the ground in a vertical direction and an oblique direction separately. The camera needs to cover an identical range of measurement to the oblique LiDAR. Data consistency should be ensured by the synchronous functioning of three devices.

The vertically mounted LiDAR and the oblique mounted LiDAR are able to detect vehicle passing times via the previously provided information. The time difference is used to calculate vehicle speed, vehicle passing times are used to calculate vehicles’ lengths, and the vertical mounted LiDAR is used to ascertain the height and outline of the vehicle simultaneously.

If the trained algorithm model leads in the recognition process, the recognition of vehicle types is enabled. The diagram below displays how the processes of vehicle type recognition and traffic flow counting should be realized:

Principle of Traffic Flow Detection and Vehicle Type Recognition

1. Traffic Flow Detection Algorithm

When the LiDAR installation angle is inputted, traffic flow detection is able to be completed.

The LiDAR installation angle is represented by θ, while the output is represented by dist=L. Dist needs to be converted into vertical distance so that the demands of different installation angles are met. The algorithm H=Lsinθ can finish the counting of traffic flow in various different environments. An adaptive algorithm update can be performed according to the threshold value of distance hopping.

2. Vehicle Type Recognition Algorithm

Initially, it is necessary to obtain vehicle passing parameters such as the vehicle’s outline, length, and height, amongst other information. Vertical LiDAR can be used to measure the outline and height of the vehicle. Two LiDARs are used to measure vehicle length, using the following calculation process:

T represents the time passing between two LiDARs, and distance is represented by S=Lcosθ. Speed is shown by V=S/T, where T2 is the passing of time of the vertically mounted LiDAR, and vehicle length is shown by L=V*T2.

The vertically mounted LiDAR can obtain information pertaining to the vehicle’s outline, the LiDAR measured distanced without the vehicle is shown by D1, D(t) shows the distance measured by the LiDAR with a vehicle present, the vehicle outline vector is represented by Dist=D1-D(t), which is the vehicle type recognition parameter.

This information has been sourced, reviewed and adapted from materials provided by Benewake (Beijing) Co., Ltd.

For more information on this source, please visit Benewake (Beijing) Co., Ltd.

Citations

Please use one of the following formats to cite this article in your essay, paper or report:

  • APA

    Benewake (Beijing) Co., Ltd.. (2019, April 29). Recognition Solutions for Traffic Flow Counting & Vehicle Types. AZoSensors. Retrieved on May 23, 2019 from https://www.azosensors.com/article.aspx?ArticleID=1626.

  • MLA

    Benewake (Beijing) Co., Ltd.. "Recognition Solutions for Traffic Flow Counting & Vehicle Types". AZoSensors. 23 May 2019. <https://www.azosensors.com/article.aspx?ArticleID=1626>.

  • Chicago

    Benewake (Beijing) Co., Ltd.. "Recognition Solutions for Traffic Flow Counting & Vehicle Types". AZoSensors. https://www.azosensors.com/article.aspx?ArticleID=1626. (accessed May 23, 2019).

  • Harvard

    Benewake (Beijing) Co., Ltd.. 2019. Recognition Solutions for Traffic Flow Counting & Vehicle Types. AZoSensors, viewed 23 May 2019, https://www.azosensors.com/article.aspx?ArticleID=1626.

Tell Us What You Think

Do you have a review, update or anything you would like to add to this article?

Leave your feedback
Submit