Identifying the location of objects with computer vision
Use computer vision techniques to determine the location of objects in a frame or scene. Object localization, sometimes referred to as ‘selective search’, is the process of using bounding boxes to identify the presence of objects in an image and indicate their location. Not only can these ‘axis-aligned’ bounding boxes indicate the position and scale of an object, they can also – when combined with other rules – determine the relative proximity of one object to another. Popular applications include:
Manufacturing
Count objects in specific areas – and combine with exception reporting – to monitor for unexpected occurrences in production (eg: missing or out-of-place objects on a conveyor belt).
OH&S
Determine the location of multiple objects, in relation to each other, to enable proximity alerts (eg: When people get too close to machines and vehicles on work sites).
Security
Automatically identity when people enter a protected zone, breach a surveilled perimeter, or pinpoint a specific object in a crowd.
Occupancy
Programmatically count the number of people in a defined region to efficiently manage venue or area capacity limits.
Retail
Help determine the popularity of shopping center amenities, stores and displays.
*Disclaimer: Outcomes described and depicted are illustrative in nature. Learn more.