0

I'm working on a project in which images/video will need to be processed in order to detect certain objects within the frame. For this project, I'm in charge of selecting hardware (image sensor, lens, processor, etc), and I'd like to select the lowest cost image sensor possible for cost-constraint reasons. For the application, the camera will adhere to the following two requirements:

  • The camera will be at a fixed position
  • The distance to the object of interest is known

So, I'm wondering if there is any data/recommendations like the following:

"Based on algorithm X (e.g. face detection), select an image sensor / lens combination so that the object of interest is covered by a pixel-density of at least 40 pixels-per-foot."

For a few use cases, such as face detection or face recognition, I've found some materials available online that have recommended minimum PPF requirements to ensure acceptable data is fed to the algorithm (Pixel Density Reqs for Face Detection). Obviously, any detection requirements are heavily algorithm dependent, but I was hoping maybe someone could provide some insight or resources they've come across to address this.

Some (slightly) more specific information - if possible, I'd like to be able to perform the following:

  • Perform facial detection (not necessarily recognition)
  • Discern/detect a human from the front, back, and top

Also, I'm aware that there are image processing libraries available for computer/machine vision (such as OpenCV), could this library (or similar ones) contain this information? I'll look into it, but if someone has a specific library to reference that would be very helpful.

Thanks!

MandM
  • 3,043
  • 4
  • 29
  • 52
  • 2
    WOW, this is a completely nebulous set of requirements. Pixels per foot AT THE OBJECT you are trying to measure actually is a function of distance from the measurement device (i.e. as they are farther from the device, they will take up less of the field of view, and thus less pixels)!!! As another addition, when something like detection algorithms are concerned, sensor noise becomes a really large problem when the object you are trying to detect is a small portion of the FOV. Thus the cheaper you go, the higher the noise (usually). – trumpetlicks May 14 '14 at 15:25
  • 1
    It depends on how big your feet are. – Paul R May 14 '14 at 15:36
  • WOW, sorry that caught you so off guard trumpetlicks. I understand both of these points and I realize my example scenario wasn't quite worded correctly. The example has been updated. – MandM May 14 '14 at 15:37
  • Didn't catch me off guard so much as it seems that the requirements that YOU have to work with aren't well defined, that's all :-) There is much more than algorithm dependence here, there is physical dependence. The lens setup will essentially define you FOV, how much actual light you will be able to gather in specific situations, which will feed into how much noise the sensor may feed to your algorithms. The FOV if static may work for one range of distances from the sensor, but not others. You are going to need range and condition requirements as well as others to solve this problem. – trumpetlicks May 14 '14 at 15:58
  • I made sure to note that the camera was at a fixed position and that the distance to the object of interest was known - this solves the "range condition requirements" you're after. Also, I'm aware of how the lens and image sensor work in conjunction with one another, but that is outside the scope of the question. I appreciate your comments, but the question is still valid - I'm not asking how to design the physical setup, I'm wondering if there are recommendations for the characteristics of the setup (there's already data for facial detection, is that data available for other use cases?) – MandM May 14 '14 at 17:25

0 Answers0