English | 繁體中文

This is a simple gesture recognition program for the MeArm. It uses the MediaPipe to detect hand gestures and control the MeArm.

https://www.youtube.com/watch?v=TBRi6ecgQfc

Servo define like this picture.

Change the initial position of the servo motor in the code.

int default_angle[4] = {75, 90, 90, 60};

make sure remove the bluetooth module from serial pin then upload the code.

cd python
pip install -r requirements.txt

you can use the normal webcam or an android phone as a webcam through Droidcam app.

change the cam_source in the code.

cam_source = "https://192.168.1.100:4747/video"
# 0,1 for usbcam, "https://192.168.1.165:4747/video" for webcam

First, you need to know the servo movement range of your MeArm.

Then, change the configuration in the python code.

x_min = 0
x_mid = 75
x_max = 150

y_min = 0
y_mid = 90
y_max = 180

z_min = 10
z_mid = 90
z_max = 180

claw_open_angle = 60
claw_close_angle = 0

Sencond, make sure the debug mode is on.

debug = True

Due to different camera viewing angles and resolutions, it may need test and then change the values of the following parameters.

# use angle between wrist and index finger to control x axis
palm_angle_min = -50
palm_angle_mid = 20

# use wrist y to control y axis
wrist_y_min = 0.3
wrist_y_max = 0.9

# use palm size to control z axis
plam_size_min = 0.1
plam_size_max = 0.3


fist_threshold = 7

make sure the com port is correct.

ser = serial.Serial('COM4', 115200)

change debug mode to False.

debug = False

run the code.

python main.py