I am looking to build a camera trap for wildlife viewing, able to capture pictures (or even videos) in decent quality to be able to identify distinct features. I do however need to be able to see the pictures in real time, so the camera would need some sort of wireless connection to a receiving device such as a phone.

I thought of using an FPV setup as used in drone racing, but the quality and range seems to be quite low, and the strengths of an analog signal seem to be latency and fps, which are both not important to me. It would be nice to have some sort of digital signal to be able to do some editing, such as storing pictures. It would also be cool to connect several cameras to one device.

Are there any solutions to this problem? Would a raspberry pi/arduino setup work in some way? Thanks!

  • 3
    This doesn't seem like a robotics question, perhaps the photography forum might be more appropriate. Oct 19 '19 at 12:32
  • 1
    @PeterCorke For the photography aspect, but the vision system and programming part aren't really in this community's primary expertise.
    – mattdm
    Oct 21 '19 at 9:42
  • I thought this was more about the photography, since it doesn't talk about moving the camera to point at an automatically acquired target. That might be more in the Robotics wheelhouse.
    – Mark Booth
    Oct 21 '19 at 9:47

Building a Wildlife Spotting Robot is an interesting amateur project which can be handled with off-the shelf hardware like the Raspberry Pi. Before the camera gets tested in the forest, it can be utilized in the garden or in the house for making photos of the own dog. The bottleneck in the project isn't located in the hardware, but in the software side. After booting the microcontroller into an operating system, the user has to decide which kind of software he likes to run in the background. In most cases, it's a combination between a webcam-capture tool plus an interval based shutter.

A simple starting point for the setup would be to capture each minute a single picture. If the sd-card in the robot-camera is full ,it gets copied to the desktop PC. A more advanced setup would use an image detection program like OpenCV to recognize the content of the picture. Only if a movement is available, the shutter gets activated.

Programming the shutter will become the core task of the overall project. The Python programming language makes it easy to script actions and create if-then-rules. One option would be, that if the background subtraction algorithm has detected a movement, the record mode gets activated which means to record a picture every second. If the movement is no longer detected, the camera robot is going into the normal mode, which means, that no image is captured.

Not the answer you're looking for? Browse other questions tagged or ask your own question.