Video object detection and depth recognition for hapwrap technology.
This program uses two different Machine Learning Models for depth perception. Within the hapticVision main folder create a new folder called "models". Download both Monodepth2 and ImageAI and place it in the newly created models folder.
Click Here To Download the Models
1.Monodepth2 - Used for depth perception https://github.com/nianticlabs/monodepth2
2.ImageAI - Used for object detection https://github.com/OlafenwaMoses/ImageAI
Python 3.5.1 or later
TensorFlow 1.4 or later
pip3 install tensorflow
OpenCV
pip3 install opencv-python
pip3 install cv2
numpy
pip3 install numpy
matplotlib
pip3 install matplotlib
PIL
pip3 install PIL
ImageAI
pip3 install ImageAi
Pytorch
Keras
pip3 install Keras
Run the main python method. Place the wanted video inside the video folder and change the video_name variable to the name of the video.
Program outputs the predicted depth, distance, height and width, and angle from camera of detected objects. Note, the program is slightly innaccurate for outputs within 10 feet of the Camera.