Danger Zone Assessment: Case Study of Edge AI Development in Video
Danger area determination using pose (human skeletal recognition) technology based on deep learning.
This is a danger area determination system that utilizes pose recognition technology (human skeletal authentication) through deep learning. It estimates a person's position from images captured by a camera and triggers an alarm when they approach a pre-set danger area. The POSE technology can only make determinations in a plane (2D), but by also using depth estimation technology through deep learning, it has enabled determinations in three dimensions (3D). It compares the pre-set danger area with the position of a person's chest in real-time, changing the chest marker from green to red in the case of a danger area. While the area outside the door is designated as a danger zone, it is also possible to specify a certain depth, such as only when passing through the door. ■ For inquiries Marble Inc. Embedded Product Division, Product Business Department Phone: (03) 3419-6011
basic information
Introducing Edge AI development examples in video What Edge AI can do: Motion Detection It detects movements and identifies people or objects while obtaining location information. Applications include detecting the movement paths of factory workers and issuing warnings when approaching dangerous areas. State Recognition It judges shapes and appearances from images. This enables shape judgment during product assembly and state assessment of cooking levels in food manufacturing processes. Classification and Identification It classifies and identifies people or objects. This includes classification based on shapes and colors learned from images, as well as trend analysis using the results for parts, boxes, etc. We are developing a demo of Edge AI based on this technology. For inquiries: Marble Inc. Embedded Product Division, Product Business Department Phone: (03) 3419-6011
Price range
Delivery Time
Model number/Brand name
Edge AI Development Case: Hazard Area Determination System
Applications/Examples of results
Please contact us.
