Vehicles as diverse as their drivers and habits

Commercial vehicles of professional workers and tradespeople pose an interesting challenge to any detection algorithm. They come in widely varying sizes, shapes, and appearances: from typical passenger cars to delivery vans, pickups, flatbed trucks, box, and curtain-side trucks, both with and without an assortment of trailers, with and without additional superstructures such as roof racks.

Parquery detects any vehicle, even motorbikes

Parquery detects any vehicle, even three-wheelers

Parquery detects any vehicle, even tiny ones

Parquery detects any vehicle, even collection cars

Parquery detects any vehicle, even sportscars

Parquery detects any vehicle, even pickups

Parquery detects any vehicle, even cars with roof tents

Parquery detects any vehicle, even cars with bike racks

Parquery detects any vehicle, even vans with racks

Parquery detects any vehicle, even large vans

Parquery detects any vehicle, even vans with ladders

Parquery detects any vehicle, even lawn mowers

Parquery detects any vehicle, even cargo net trailers

Parquery detects any vehicle, even earth moving vehicles

Parquery detects any vehicle, even trucks

Parquery detects any vehicle, even trucks at night

Parquery detects any vehicle, even truck trailers

Parquery detects any vehicle, even trucks and trailers

Parquery detects any vehicle, even coaches

Parquery detects any vehicle, even buses at night

Parquery detects any vehicle, even campervans

Parquery has seen it all, small, medium, and big: motorbikes, 3-wheelers, tiny motorcycle-cars, oldtimers, sports cars, pickups, vehicles with roof-tents and bike racks, delivery vans with and without roof racks, delivery vans of any size, shape, and color, trailers with and without cargo, earth-moving machines (at the very edge of the camera image), large trucks, road tankers, buses and coaches, camper vans, ...

Parquery even detects shapeshifters

Furthermore, a vehicle's appearance can drastically change depending on its cargo: open freight platforms of trucks, pickups, or trailers may be empty or covered with materials and equipment. A gardener's truck, for example, might look like a treetop when loaded with shrubs. A trailer laden with shopping carts is still a vehicle and should be recognized as such, contrary to shopping carts in an empty bay. The cargo - and with it, the vehicle’s look - might even change over time while the vehicle is parked.

Cargo can range from palettes, boxes, tools, appliances, wooden boards, bushes, trees, over sand and gravel, to construction and earth-moving machinery such as excavators. Ladders, wooden boards and beams, equipment, and bits and bobs are also frequently strapped onto roof racks. All of these profoundly change the characteristics and appearances of the vehicles and challenge any detection algorithm.

Parquery detects any vehicle, even with empty flatbeds

Parquery detects any vehicle, even with flatbeds

Parquery detects any vehicle, even with stuff on flatbeds

Parquery detects any vehicle, even with machines on flatbeds

Parquery detects any vehicle, even with various equipment

Parquery detects any vehicle, even with heterogeneous flatbeds

Parquery detects any vehicle, even with uneven flatbeds

Parquery detects any vehicle, even with unevenly shaped objects

Parquery detects any vehicle, even with boxes on flatbeds

Parquery detects any vehicle, even with large ladders

Parquery detects any vehicle, even with vehicles on flatbeds

Parquery detects any vehicle, even with vegetation on flatbeds

Flatbed vehicles with and without cargo and cargo nets present intriguing detection difficulties due to their diverse appearances.

Why does it work so well?

Dependable in cluttered situations

Parquery’s algorithms can be trained on any vehicle - as diverse as motorcycles, trains, earth-moving machinery, or boats - while being robust to visual clutter or other objects not relevant for detection. Neither open doors and trunk lids of vehicles being loaded or unloaded, nor shopping carts, pedestrians, or bicycles in the foreground interfere with recognition. Parquery’s AI can reliably distinguish oil stains on the asphalt, patched-up road surfaces, or litter from a parked vehicle.

Parquery does not get confused between shopping carts and vehicles

Parquery does not get confused between trolleys and vehicles

Parquery does not get confused between pedestrians and vehicles

Parquery does not get confused when doors are open

Parquery does not get confused when people are on the parking spot

Parquery does not give inaccurate results if pedestrians are on parking spots

Parquery does not get confused between stuff and vehicles

Parquery does not get confused when bikes are on parking spots

Parquery does not get confused when cyclists are on parking spots

Parquery does not get confused by tyre marks or wet areas

Parquery’s system is robust to occlusions and other objects such as trolleys, car doors, pedestrians and cyclists passing through, oil stains, wet patches, puddles, or tyre marks.

Beyond cars and trucks

Parquery’s intelligent algorithms learn by example and experience and, thus, can be trained to detect and recognize any object. Beyond solving smart parking problems, Parquery also works on people counting (footfall data), traffic monitoring, and warehousing solutions.

Have you got a project to challenge us?