I am working on a project where I need to detect yellow junction boxes on the road from a vehicle’s dashcam feed. I am considering two approaches:
Edge Detection: This method might not be reliable if the junction boxes are unclear or partially wiped off.(Attached Images -Unclear and Unclear2)
Training a Custom Object Detection/Segmentation Model: Using a model like YOLO trained on sample images of yellow junction boxes.
What would be the best approach to solve this problem?
[Ideal](https://i.sstatic.net/65d51uhB.png)
[Unclear](https://i.sstatic.net/4awCUSXL.png)
[Unclear2](https://i.sstatic.net/M6t5Rh6p.png)
Rohan Bhavale is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.