This repository contains the implementation of the YOLOv3 algorithm for the calculation of freely falling object movements. The steps involved in the process are outlined below:
-
Training Model using Darknet Framework
- Train the YOLOv3 model using the Darknet framework developed by Joseph Redmon.
- Utilize the provided pre-trained model for the base architecture.
-
YOLOv3 Architecture Overview
-
Loss Function
-
Mean Average Precision (mAP)
-
Object Tracking Results
-
Time Measurement Results
-
Acceleration Calculation Results
- For accurate results, ensure the YOLOv3 model effectively detects both the falling object (ball) and the sensors.
- The architecture of YOLOv3, along with its training and loss function details, play a pivotal role in obtaining reliable outcomes.
- The object tracking results should demonstrate the capability of the model to precisely locate the falling object and sensors.
- The calculated time measurements serve as the foundation for determining the acceleration due to gravity.
- Comparative analysis between YOLOv3-based measurements and sensor-based measurements offers insights into the model's accuracy.
Feel free to explore the repository and refer to the attached files for detailed information about each step and the corresponding results.
This project is licensed under the MIT License.