Edge Detection vs Motion Detection
Traditional motion detection and computer vision use very different processes. Traditional motion detection is looking for pixel changing colors in a grid. Computer vision uses edge detection, which determines the outline of shapes and can tell difference between the foreground and background.
Object Detection, Image Classification, Semantic Segmentation
Computer vision can do three main things:
(1) Object detection: which answers "is a particular thing here?" and "if so, where in the image is it?"
(2) Image classification: answer what sort of subset something is, so if the object found is a "vehicle", a classifier might filter those vehicles into subgroups like "SUV" or "Truck"
(3) Image segmenter: which outlines a perfect mask of the object area, such as "the road" or "the grass"
Here's what you can do with all these new powers:
Footage that matters
Cut through the noise. Find what you are looking for, quickly and easily.
Sort videos by detection type, location, system, tag (a description of the camera, such as "outside"), camera name, date and time.
Pull Stats on Activity
Find out what moves your needle. Understand when Cars or People increase at your location, on a daily or hourly basis.
View Activity on a 24/7 Timeline
Scroll over the events and see snapshots of what the AI found.
Understand Activity on a per-camera basis.
Know when People or Vehicles are in video streams.
Get alerts via email or text message.
No more alerts when the wind blows.
Not all areas are of concern
Easily exclude areas from detections and stop getting alerts about your neighbor's activity
You may also be interested in our in-front-of / behind fence classification model and semantic segmentation models in our model zoo.
Survail's Computer Vision and AI Models
The first thing that you probably want to do is discover what computer vision analysis you can perform with survail, and so you will want to look through our availible computer vision models. Survail is a platform with a constantly growing list of objects it can detects and details it can infer from those objects.