YOLOER stands for You Only Look Once and Estimate Range while V1 implies the version of object detector (V1 = YOLOv5). The project is about REAL-TIME Object Detection and Distance/Depth Estimation using the YOLOv5 object detector. To test the Real-Time functionality immediately, we provide the option of running the model on the webcam. Moreover, the options of testing on recorded videos and images are also provided.

In order to test any Real-Time system, the most convenient method is to run on the webcam. So, we provide the options of quick inference on a local machine and visualization through the webcam. Some installations are required before running the inference and the following subsection contains the entire method. So follow step by step.

Moreover, we prefer working in Conda environments and it is recommended to install it first. In case of not have Conda installed, just skip the Conda-specific commands and follow along.

  conda create --name YOLOER_V1 python=3.7 -y && conda activate YOLOER_V1 

  git clone https://github.com/HassanBinHaroon/YOLOER_V1.git

  cd YOLOER_V1/object-detector_PLUS_distance-estimator

  pip install -r requirements.txt

  python setup.py

  python detect.py --save-txt --weights yolov5s.pt --conf 0.4 --source ../videos/car.jpg --model_dist [email protected] --weights_dist [email protected]

Click on the following link.

https://colab.research.google.com/github/HassanBinHaroon/YOLOER_V1/blob/master/YOLOER_V1.ipynb

Note! The project is still in progress.