Computer Vision Robot
Library for computer vision code in Google Colab and Raspberry Pi. Uses Yolo v3 with OpenCV.
Config:
Place the following files in the same directory as your script to configure.
- CONFIG_CV_SERVER.py - runs in server mode
- CONFIG_CV_ROBOT.py - runs in server mode on robot
- CONFIG_CV_LIVE_IMG.py - replaces server preview with matplotlib window
By default, the code will not use cameras (great for Google Colab!)
Example:
#Imports
from CV_Robot import robot
from CV_Robot import vision
from time import sleep
#Example - find objects in an image
img = vision.load_image("sample_images/sample_005.png")
vision.show_image(img)
vision.show_objects(img, pause=True)
print(vision.find_objects(img))
#Example - drive forward until a STOP_SIGN is found
robot.forward()
vision.activate_camera()
while True:
img = vision.get_camera_image()
if img is None: #This is good practice - it will stop the loop if the video ends / the camera fails
break
vision.show_objects(img)
if vision.Objects.STOP_SIGN in vision.find_objects(img):
break
sleep(0.1)
robot.stop()Data Sources
https://github.com/nandinib1999/object-detection-yolo-opencv
#Download some sample videos + images for testing
wget https://github.com/nandinib1999/object-detection-yolo-opencv/raw/master/videos/pedestrians.mp4
wget https://github.com/nandinib1999/object-detection-yolo-opencv/raw/master/videos/car_on_road.mp4
wget -O sample_images.zip https://github.com/RobertJN64/CV_Robot/raw/master/Examples/Samples/100samples.zip
unzip sample_images.zip -d sample_images
