ai-hik-tracking

Control DMX Laser using Deep Learning object detection

Playing Card Control keys

Key	Effect
Num 1-9	Change YOLO Profile
Ctrl+Num 1-9	Change YOLO Profile and backend between darknet / OpenCV
p	OpenCV DNN toggle FP32 / FP16
Ctrl+s	Save current frame with recognize result
Ctrl+c	Toggle using CPU / GPU preprocess
b	Preview split / black result
z	reduce preprocess alpha 0.05
x	increase preprocess alpha 0.05
c	reduce preprocess beta 5
v	increase preprocess beta 5
g	reduce drawing approx thresh 0.001
h	increase drawing approx thresh 0.001
F1	Change split analyize mode

Shared Control keys

Key	Effect
Num 1-9	Change Palcon-D PTZ Preset
Shift+Num 1-9	Save Current Position to Palcon-D PTZ Preset
U/D/L/R	Cal Mode: Move Laser
i/j/k/l	Cal Mode: Move PTZ
Ctrl+z	Pop last cal result
+/-	zoom in/zoom out laser
Escape	Switch between cal / test mode

Training Data Refereence

• https://github.com/okmd/playing-card-dataset.git

Use components

• https://github.com/magicbear/py-hikevent

Data Prepare Toolkit

• https://github.com/spytensor/prepare_detection_dataset.git

Project Reference

• https://github.com/AlexeyAB/darknet

• https://github.com/facebookresearch/detectron2

• https://github.com/EdjeElectronics/OpenCV-Playing-Card-Detector

• https://github.com/a-arbabian/playing-cards-object-detection