Camera Controlled Robot Arm - Adrian Bartoš

Project Goal

Control of 6DOF robot arm with camera using openCV. Robot is going to pick object and place it on oposite side of working area where coordinates are mirrored.

Parts Used

• Arduino Uno
• Servo Driver PCA9685
• 6x servos TD-8135MG
• Mobile Phone for Camera detection

Assembely

1. Build robot arm controled by arduino
2. Set up Camera detection
3. Move robot acording to coordinates gathered from camera

Camera Detection

I chose to use OpenCVP's built-in AruCo markers for defining working space and desired object.

1. We need to generate markers
   • In code with name "GenereteArCo.py" we need to genere 4x 4x4 Aruco markers and 1x 6x6 marker for object detection. For working space AruCo's we need to generte id's 1-4. So we are only changing id variable. And then depending on camera distance and recording quality the tag size variable for size in pixels. So for 4x4 Aruco marker with id 1 and size of 100px we modify code like in example below. Same thing we do for object AruCo, we change form 4x4 to 6x6. I chose different AruCo type for object for easier recognition. When all markers are genereted. We need to print them.

aruco_type = "DICT_6X6_50"
id = 1

arucoDict = cv2.aruco.getPredefinedDictionary(ARUCO_DICT[aruco_type])

print("ArUCo type '{}' with ID '{}'".format(aruco_type, id))
tag_size = 100
tag = np.zeros((tag_size, tag_size, 1), dtype="uint8")
cv2.aruco.generateImageMarker(arucoDict, id, tag_size, tag, 1)

Photos