After 6 months learning program language such as Python, Java, JavaScript, HTML, JQuery, CSS, Nodejs, MySQL, etc.. , my team came out with this project: self-driving car with driver drowsiness detection and self parking. I was in charge of self-driving model training and tuning, Donkey Car resource code modifying, camera streaming and MQTT communication set up.
21. 網頁 & Rpi整合流程
Rpi
fdetect_flask.py
常駐程式 run
PC
JAVA Web run
Rpi open
pi-camera
web publish,
say “go”
to Rpi
Rpi flask
streaming
iframe
streaming to
web
web publish,
say “takepic”
to Rpi
Rpi subscribe,
say “takepic”
take picture
and run
recognition function
recognition result
compare result
with users’ DB
in mysql
Rpi publish
say “yes”
Rpi publish
say“no”
web dispatch
to home page
web dispatch
to login page
張恩嘉 Kathy
功能網站團隊介紹 專題架構 臉部辨識 疲勞示警 自動駕駛 閃避停車 意外通報 總結 未來展望
22. MySQL
PC WEB Rpi
Email login Face Recognition login
張恩嘉 Kathy
功能網站團隊介紹 專題架構 臉部辨識 疲勞示警 自動駕駛 閃避停車 意外通報 總結 未來展望
43. 模組訓練
Load config(type, verbose, …)
↓
Load data (picture & json file)
↓
Shuffle data and split the data
↓
Generate train & test data
↓
Estimate steps for each epochs (gradient descent) & validation
↓
Define epochs & rule of early stop
↓
Train the model with Convolutional Neural Network
↓
Visualize the result & save the model
邵柏元
閃避停車團隊介紹 專題架構 臉部辨識 功能網站 疲勞示警 自動駕駛 意外通報 總結 未來展望