Stay time and car park use, measured at every gate - Casey Complex, City of Casey

City of Casey needed to know how long vehicles stayed at Casey Complex and how full each car park ran, across a site with many ways in and out. XVision measured both, with 96.8% plate detection in the field and a live dashboard Council could use and export.

Client	City of Casey, Victoria
Project	Casey Complex Vehicle Tracking and Car Park Monitoring
Site	Casey Complex, New Holland Drive, Cranbourne East VIC 3977 - Casey RACE, Cranbourne Library and Casey Radio, Casey Stadium, The Factory and the construction car park
Delivery	Temporary hire deployment, March to June 2023
Most reliable data window	26 April to 24 June 2023
Sensing	10 units: 3 plate-reading (LPR) units at the complex entrances and 7 HD occupancy units at the car park entrances
Outputs	Multi-camera dwell time, live occupancy dashboards, training, CSV export
Headline result	96.8% plate detection under real-world conditions

At a glance

• 96.8% plate detection measured in the field at Casey Complex.
• Most visits were short — median stay 1 hour 6 minutes, and 81.7% of stays under 2 hours.
• Stadium ran hottest — peak 395 occupied spaces (median daily peak 294).
• Construction stayed light — peak 37 against Race 301 and Stadium 395, because works access was measured on its own.
• Data Council owns. Occupancy in 15-minute intervals and average dwell in 5-minute intervals, exported as CSV and packaged by Council for its open data platform.

The challenge

Casey Complex is a busy community precinct in Cranbourne East. Leisure, library, radio, stadium and event users share one site. At the time, a construction car park also operated inside it. Vehicles enter and leave from several roads, and when the marked bays fill, cars spill onto grassed overflow areas.

Council's Growth and Investment team needed hard evidence on three questions:

1. How long do vehicles stay on the complex?
2. How full is each car park, through the day and across the week?
3. How much parking demand sits outside the marked bays?

Manual counts give a snapshot of one day. Staff cannot watch every entrance every hour. The construction access also had to be measured on its own, so construction parking did not inflate the visitor overflow figures.

Council needed a continuous record of stay time and car park use for the whole complex, without a team watching screens.

The solution

XVision designed the deployment around two layers of evidence: who enters and leaves the complex, and how each car park fills.

Layer 1 - Plates at the complex entrances

Three plate-reading (LPR) units cover the complex entrances on Main Street, Race Drive and New Holland Drive. Each unit reads the number plate of every vehicle entering and leaving. XVision matches each plate across the entrance units, so stay time comes from when a vehicle arrived and when it left - not from a count crossing a line.

Layer 2 - Occupancy at each car park

Seven HD units cover the car park entrances: Casey RACE, the Cranbourne Library and Casey Radio car park (main and side entrances), Casey Stadium (north-east and north-west entrances), The Factory and the construction car park. Each unit counts vehicles in and out, which gives live occupancy and percentage full for every monitored car park.

Vehicles that enter the complex but are not counted into a monitored car park are attributed to grassed overflow. That gives Council a measure of overflow demand without extra hardware on the grass.

Processing and data path

Plate reading and vehicle counting run on the device at each unit. Only the results - plate matches, entry and exit events and counts - travel over 4G to the XVision platform. All units sit on semi-permanent poles chosen during a site inspection, and came down at the end of the hire.

Dashboards and exports

• Live dashboard. Availability and percentage full for each car park, plus stay time for the complex. XVision trained Council users on the dashboard during the study.
• CSV export. Occupancy per car park in 15-minute intervals, and average dwell time in 5-minute intervals from the entrance units.
• Council-ready format. Council asked for CSV and spreadsheet data with locations rather than a narrative report. XVision delivered to that format, with unit locations for mapping.

Installation QA

XVision checked every unit at commissioning. All 10 units were active and passed QA: connectivity, communication and counting.

The results

96.8% plate detection in the field

Plate detection at Casey Complex measured 96.8% under real-world conditions, on live traffic at the complex entrances. This is the figure that makes stay time trustworthy. When nearly every plate is read at entry and again at exit, the dwell record reflects how the complex is actually used.

How the 96.8% was validated. Validated over 26 April to 24 June 2023 across 18,310 dwell-time records from the three plate-reading units, achieving a 96.8% plate detection rate.

Most visits were short

Across 18,308 cleaned dwell records in that same window, median stay was 1 hour 6 minutes. 81.7% of visits lasted under 2 hours. Stay time for the complex is therefore a short-visit pattern, not a long-stay default — evidence Council can use when setting time limits and planning bay mix.

Stadium ran hottest

Of the visitor car parks, Stadium carried the highest pressure: peak 395 occupied spaces, with a median daily peak of 294. Live occupancy showed which car parks were under pressure and when, so Council could see Stadium as the clear peak demand site without relying on a one-day manual count.

Construction stayed light against visitor parks

The construction car park peaked at 37 occupied spaces over the study window, against Race 301 and Stadium 395. Measuring works access on its own kept construction demand out of the visitor record — the design choice paid off in the numbers.

Occupancy validated against manual counts

Occupancy counts were checked against manual counts on site in the final weeks of the study. This check covers bay occupancy only and is separate from the 96.8% plate detection result. RACE, Stadium and Construction car parks matched the manual counts closely. Where the checks found an issue - an in/out direction label on two units - XVision corrected it and reapplied the correction to the full historical record, so every occupancy figure Council received matched the raw entry and exit data. During data validation, the Library car park unit was underperforming and needed an aim adjustment to return to the promised accuracy.

Evidence Council kept

The hire units came down at the end of the study. The evidence stayed with Council: the stay time and occupancy files, unit locations, and a 26 April to 24 June 2023 record that Council packaged for its open data platform. Council used the evidence to guide parking planning for Casey Complex.

What it means for councils and venue operators

Multi-entrance precincts - leisure centres, sports and event venues, library and civic hubs, hospitals and campuses - share the same problem. Counts at one gate do not explain how a site is used. Stay time and car park occupancy together do.

• Measure stay time at the gates. Plate matching across entrances gives stay time for the whole site, without staff timing visits by hand.
• Measure occupancy at each car park. Live percentage full shows which car parks are under pressure and when.
• Account for overflow. The difference between complex entries and car park entries shows demand outside the marked bays.
• Separate what should be separate. A construction or staff access measured on its own keeps the visitor record clean.
• Keep the data. CSV exports at fixed intervals go straight into Council's own analysis, GIS and open data work.
• Deploy and remove cleanly. Temporary units on semi-permanent poles collect a defined study period, then come down.

Your engineers review the evidence and make the decision. XVision gives them a record they can defend.

About XVision

XVision helps traffic and transport engineers see how people and vehicles actually behave, and turn that evidence into a decision. XVision combines specialist sensing with the road and parking data councils already hold, so teams can review the evidence and decide the next step in minutes, not days. XVision has delivered 250+ deployments across 7 countries.

XVision AI is a trading name of Metamorph Business Pty Ltd.

Reference

Chris Williams, Smart Cities Project Manager, City of Casey

Discuss a location

Tell us the site, the stay time and parking questions you need answered, and the decisions they support. Could we spend 15 minutes in the next two weeks discussing one site?

Simon Maselli - [email protected] - xvision-ai.com