Traffic congestion increases the time required to commute. We all know this too well here in the bay area, like any urban citizens across the world. It inflicts increased operational costs on the urban transport system. Many forecasts suggest that this will only get worse in the years to come.
This rise in congestion has pushed governing authorities to promote the usage of public transport vehicles instead of private ones. Few cities have also tried building dedicated roads for public vehicles and few others have tried to lure commuters by lowering the costs.
Even after several measures, a most common issue arising with public transport is that they are overcrowded during peak hours. Many people for this sole reason abandon the public transport and use alternative ways to commute in peak hours. While many cities are aiming to be transformed as smart cities by the end of this decade, public transport congestion poses a genuine challenge to the smart infrastructure.
Overcrowding of public buses is caused due to the negligence of public authorities in managing the load on every single bus. The load on each bus on the same route at a given time window varies since the bus which arrives first to the terminal predictably sees a greater load.
Since passengers lack real-time information on bus schedules and their current load, they assume that the next bus arriving at the terminal has to be inevitably boarded. If they can get a real-time data about the crowd density in each bus on a real-time basis on their mobile app, they can plan their travel accordingly.
Estimating the exact crowd density in each bus can be achieved by implementing people analytics coupled with high-resolution 4K camera input. That being said, it requires a set of cutting-edge technologies to make the accurate estimation.
Following are the camera technologies which augment the efforts of reducing the in-bus congestion in a smart city ecosystem.
Green Polygon Mapping the Congestion Area
Polygon mapping is a technique of storing spatial information in a closed polygon loop in the form of data objects. The enclosed area will also include the boundaries which form the closed loop. A green colored boundary is superimposed on passenger compartment of the bus which has to be analyzed for a number of human heads in it.
Then the three following aggregation techniques are applied to that area to determine the exact number of passengers in the green polygon.
Congestion Heat Maps
One of the ingenious technology to determine interior crowd positioning is ‘collaborative crowd sensor’ using heatmaps. It can automatically detect crowd density, crowd movement, and direction of the movement in the defined green polygon loop. Apart from estimating the crowd-density in a certain area or overall area, it can also capture exposure time of the specific pattern. Crowd patterns are recorded and then analyzed. The analyzed data is sent for contour mapping to draw deeper insights on crowd density.
Contour Mapping
The contour map is a topographic map that determines the shape of an object by perceiving its physical features like curves and contour lines. Using functional imaging technique, the number of head contours in a specific polygon mesh is calculated. Using this, the number of people in the bus is estimated.This data will be updated on a real-time basis to the central server.
Contrast Mapping
The occupancy analysis is done using contrast mapping with the help of real-time video feed sent by these high-resolution cameras. Here the occupied and unoccupied spaces are calculated.
Contrast mapping is a technology which can differentiate between objects with the help of their reflective indices. Two objects are considered different if there is immediate shift it the contrasts of a color spectrum.
The video captured from high-resolution 4K cameras will have predefined contrast values to determine the presence of a human being on the bus. For example, the contrast range of the skin complexion, hair, bald head and background color are initially hard coded. This data is compared with the estimation we had through contour mapping. By analyzing both of these data, the exact number of people in a specific bus is calculated.
Three wide-angle camera feed connected to a central server will parallelly function to have different views of the crowd. To achieve the precise count of passengers in each bus, we need high-quality video feeds which can fetch sufficient information during different climatic variations and also in low lighting conditions.
A 4K UHD camera will help us in achieving high-quality video feed. By coupling a 4K video feed with AI-based algorithms, we can send real-time occupancy updates to the passengers thereby reducing the in-bus congestion. A well-managed transport system helps us to make smart city possibilities more achievable.
Last Note: Even in the cities where bus congestion is not a challenge, this technology can be applied to increase the ROI of the transport. Many urban buses witness less than 40% occupancy even at peak hours. By having a real-time crowd estimation system integrated with schedule and dispatch, we can reduce the frequency of buses with the help of AI-based prescriptive insights.
