Announcement

Collapse
No announcement yet.

Infrared Room Occupancy

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    Infrared Room Occupancy

    This is a long-discussed topic. There is a possibility to evaluate occupancy in a room with an infrared camera, a pi, and enough skill in software programming.

    https://blog.platypush.tech/article/...chine-learning

    Personally I managed to mount the camera on the Pi, get plenty of photos, webserver running, but I was unable to continue with the jupiter notebooks to get the training phase and deploy it on the pi. The idea is simple, the pi obtains an image every 30 seconds and compares it with the training images, send a curl - post to homeseer that keeps a virtual device of room occupancy on or off.

    Hardware I tried:

    Infrared camera US 65 : https://es.aliexpress.com/item/10050...&pdp_ext_f=%7B %22sku_id%22%3A%2212000026428105129%22%7D&pdp_pi=-1%3B38751.0%3B-1%3B-1%40salePrice%3BCLP%3Bsearch-mainSearch

    Pi-4: US 120

    Someone with more software skills help to resurrect this project?

    #2
    This might be useful.
    https://medium.com/@epraveenns/real-...n-b2f541accc8e
    Wade

    "I know nothing... nothing!"

    Comment


      #3
      I've study this topic for several years now, using IR and/or imagery is not useful/desirable, depending on your point of view. First issue is privacy, some people (in your household) might object to images being continuously captured in certain rooms. The other is detector/camera angles not always being optimal and false positives.

      For large scale commercial installations the most successful occupancy detection systems use a combination of microwave, IR, and CO2 detection, and machine-learning software. It seems the average human puts out enough CO2 when entered into an area that it can be easily detected with off-the-shelf components, and consistent enough that correlating with the other detection methods mentioned, one can actually get an accurate count of living humans in the area. This solves the issue with people being very still while reading, sleeping, watching TV, etc.

      Do enough googling and there are several papers/studies done in this area.

      Comment


        #4
        You can also read about this one : https://www.synology.com/en-uk/products/DVA3221

        Eman.
        TinkerLand : Life's Choices,"No One Size Fits All"

        Comment


          #5
          Originally posted by Eman View Post
          You can also read about this one : https://www.synology.com/en-uk/products/DVA3221

          Eman.
          There is a QNAP solution as well if I recall...

          Sent from my Pixel 6 Pro using Tapatalk

          Comment


            #6
            Originally posted by Eman View Post
            You can also read about this one : https://www.synology.com/en-uk/products/DVA3221

            Eman.
            That one seems a little pricier.
            Wade

            "I know nothing... nothing!"

            Comment


              #7
              Originally posted by Sgt. Shultz View Post

              That one seems a little pricier.
              Man, you only live once! A few months down the road, you won't feel the pain... It takes care of a lot of things so...


              Eman
              TinkerLand : Life's Choices,"No One Size Fits All"

              Comment


                #8
                But the project is for a house installation, common areas, not rooms, and the thermal camera its about 768 pixels, its onlly detecting spots of heat (face, hands) and its unable to identify a person, just inform the living room its occupied. Plus the hardware is very cheap. I will try to reach to blog author.

                Comment

                Working...
                X