How can computer vision help detect the spillage of hazardous material?


Aug. 31, 2021

The potential applications of computer vision are nearly limitless: from counting the number of customers in a store to diagnosing complex medical issues. Practically any use case that involves visual data (i.e. images and videos) is suitable for computer vision, with state-of-the-art models often matching or exceeding human performance.

Detecting the spillage of hazardous material would fall under the use case of industrial AIoT (Artificial Intelligence of Things). There are several ways you could detect spillage or leakage, depending on which kind of hazardous material we’re talking about—for example, sensors that can identify certain harmful chemicals in the water or air. If we’re talking strictly about computer vision, however, then there would have to be some kind of camera or other visual surveillance system in place, as well as some kind of noticeable visual change that would indicate a spillage. This could be some kind of defect in the storage containers, a visible drip or puddle, etc.

detecting spills with computer vision AI from Chooch AI

Formally, in computer vision this kind of use case would fall under one of two subfields: event detection and/or object recognition.

  • Event detection is a subfield of computer vision that seeks to Identify when a given event has occurred. Here, the event could be some kind of damage to the storage container, or a visible leak on the floor. Computer vision models already excel at event detection use cases such as fires and falls.
  • Object recognition is a subfield of computer vision that seeks to recognize and identify the most prominent objects in an image. Here, you could try training a computer vision model to recognize a separate “puddle” object category. This would probably work best if the hazardous material in question were a colored liquid, not a clear one.
    Of course, a production-ready, enterprise-grade computer vision system would need some safeguards in place to distinguish between hazardous and non-hazardous material, or at least to indicate the potential level of uncertainty. Not every spillage incident is cause for concern—you don’t want an alarm going off every time someone spills a glass of water!

Oops, something went wrong.
Please try back later.

Click here Tap here to upload Upload your image
and Chooch will recognize it.
Supported formats are .jpg .jpeg .png