【E504】TensorFlow+OpenCV模拟自动驾驶

2021-08-23 16:42:11      索炜达电子      458     

项目编号:E504

文件大小:13M

源码说明:带中文注释

开发环境:Python

简要概述:

Self_ driving at Euro Truck Simulator2 -----《欧洲卡车模拟2》

#Implementation:

Control part: Use win32 api to control keyboard, which self play Euro Truck Simulator2.

Trace the traffic-line part: (openCV):

1, transfer original image to gray cv2.COLOR_BGR2GRAY

2, than using Canny to due with the image:

【E504】TensorFlow+OpenCV模拟自动驾驶

作用是勾勒路线的轮廓


3, using cv2.HoughLinesP() function to analyze 道路的虚线和实线 原理是计算条线之间的空隙。而实线比较好处理。 enter image description here

【E504】TensorFlow+OpenCV模拟自动驾驶

4, Using two lines to figure out "Turn", and the signal will call win32 to keyboard.


二, Use TensorFlow to recognized front of view, such as Car or Traffic_light.     This part based on :object_detection


Problem: 通过CNN训练模型,但还是未能解决红绿灯信号识别的问题,打算用openCV做颜色归纳 enter image description here   行人检测

【E504】TensorFlow+OpenCV模拟自动驾驶

行人检测

【E504】TensorFlow+OpenCV模拟自动驾驶

文件列表:

目录│文件列表:

 └ AutoDrive

    └ AutoDrive

       │ alexnet.py

       │ autoDrive.py

       │ balance_data.py

       │ directkeys.py

       │ ez_setup.py

       │ getkeys.py

       │ grabscreen.py

       │ Maching.py

       │ main.py

       │ tf_vehicle.py

       │ training_data_2.npy

       └ screenshoot

          │ 476419097427111374.jpg

          │ 547439799077907042.jpg

          │ Thumbs.db

          │ truck_Canny.png

          │ truck_light.png

          │ truck_line.png

          │ truck_trafficlight.png

          └ WechatIMG60.jpeg

TAG模拟自动驾驶
  • 9 次
  • 1 分