Hello all,
I am trying to build a system that could identify changes in images taken every day (24h delay between each image). My approach was to use backgroundsubtractionMOG2 but it's failing when the weather changes dramatically (sunny to cloudy day) and learning the background is taking a lot of time.
Do you recommend any algorithm or method to solve this?
Thanks.
What exactly do you mean by "identify changes in images"? Do you want to detect new objects that enter the scene or understand the changing weather conditions? The answer could change depend on what your goal is.
For more robustness against weather changes, you can try deep neural networks (esp. convolutional neural nets) provided that you have enough training images. But depending on the actual images, the solution could also be much easier. What are you trying to achieve with background subtraction?
Hello Kevin, Thanks a lot for your fast reply.
I am mainly trying to detect any new object in the scene. I can use a neural network to detect every single object (car, humans, etc...) but i fear it's going to be costly as this requires a lot of data. For this reason i thought of identifying the background and any change in the background will be considered as an intruder or "change".
Asked: 2017-02-16 12:10:35 -0600
Seen: 225 times
Last updated: Feb 16 '17
could you provide sample images