Take a few simple, straightforward steps to integrate Parquery's results into your existing platform or access them via a browser-based dashboard. Here's how you can make Parquery's solution work for you.
Parquery detects objects in images taken from any camera, also those already installed. All we need are the images they produce. No video stream required. Parquery can detect vehicles in video camera images at any interval the client requires, for example every 10 seconds, every 30 seconds, once a minute, every 5 minutes, etc.
If you haven’t installed any video cameras yet, Parquery will help you identify the best, most effective locations for installation. All we need is a map of your parking areas and the available installation height. Then we’ll run a simulation based on standard cameras available on the market.
Our goal is simple: to save you money by covering as many parking spots as possible using as few cameras as possible.
Any camera will do.
Parquery can use images from any Internet Protocol (IP) camera, including standard cameras, CCTV cameras, webcams and more.
Once your cameras are supplying images at the desired refreshed rate, Parquery can analyze them to detect objects and silhouettes. This analysis can be conducted locally on the customer’s premises, or remotely via Parquery’s cloud servers.
Parking analyses performed on Parquery's cloud servers require images from your cameras, so you’ll need an internet connection: ethernet cable, Wifi, 3G/4G SIM cards, etc.
Parking analyses performed locally onsite require Parquery algorithms installed on a local server. In this case, you can simply transfer images to this server via a local network or via the internet.
Parquery’s software doesn’t need any major infrastructure. Whenever our customers want to install and run our software onsite, we put them in touch with our Operations Team. Our technical experts will help you identify the best solution for your parking areas based on your existing servers.
Here is an example of server requirements for an off-street parking area with 200 spots:
Linux Virtual Machine (or specific Linux Machine);
Minimum 2 CPUs;
Minimum 4 Gb RAM;
Approximately 100 Gb of storage space on your hard drive (including space for Linux).
As soon as Parquery receives your camera images, we can deliver results. These initial results achieve 95% accuracy, and will improve to +99% accuracy within the first month.
During this period our Operations Team will train our software to adapt to your specific environment, identifying and compensating for potential edge cases including extreme shadowing, partial tree obstruction, adverse weather conditions, etc.
Prior to delivering your initial results, we’ll organize a video conference call together in order to make sure we send your data in a format that responds to your specific needs.
One month after you have received your results, we organize a second meeting to gather feedback on your experience and, if necessary, modify the dashboard to better suit your purposes.
Depending on your needs and suggestions, we’ll also use this meeting to define any next steps.
Parquery's solution is easy to implement and it is possible to use existing cameras.
It is an ideal solution for parking control, which is also more efficient than sensors (image proofs).
D. Gnerre, City of Vevey