Birdy Cam is getting ready for the Winter...
This is an ESP32 Cam being fed into MRL's OpenCV service, streaming Mpeg with URL (plus Motion Detect Filter).
If you have an ESP32 Cam then it is HIGHLY recommended to install a good antenna (see below).
For point of cost reference ESP32Cam is circa $6 (one bag of sunflower seeds also circa $6).
The images used here are UXGA(1600x1200) Streamed over WIFI through 4 concrete walls 25metres from router to balcony bird table.
For ease of use/longevity & placement a 3D printed clip and swivel mount design has been implemented.
Beneath in its holder is a 4000mAhr Lipo Battery pack (cost 4CHF in sale) which should give many hours of use.
In the GIF above you can see kwatters latest Motion Detect filter in action. It detects very well indeed for the fast moving birds... at the end of the clip you can just see it catch a bird flying by.
Also what is great to see during the Motion Detect ...is the afterburn effect of the green detect boxes, I guess relearning the static background image!!.
All that needs to be done now is find out where the Motion Detect Filter "methods" are (Hidden), and write some Python code to detect and blip the record video method (!!) for automatic archive and replay. (Thats the wish and plan).
N.B. The bird table is a 3 part 3D design that holds and secures all the bits cut from a single 30x30cm plywood board.
Example of motion detection python script.
I just added one that should work... pretty simple script/example.
it registers a callback for the publishMotionDetected method.. in the callback you get the list of rectangles that show where the motion was detected.
https://github.com/MyRobotLab/pyrobotlab/blob/develop/home/kwatters/Mot…
enjoy!