contact point detection and tracking
Budget: $30 – $250 USD
Please go through the project details and lets discuss afterward.
Here is what I am expecting:
1. Find he contact point than keep detecting or tracking the correct contact point even switching of catenary wires in video input.
2. Making a full pipeline using yolo model to detect Pantograph, converting your model or algorithm to Onnx format and running inference with opencv dnn module using c++ code.
3. Testing on different videos or scenarios and overall accuracy which should be more than 85%. If needed reducing DL Model Size, optimization for Nvidia Board and obtaining more 15FPS speed.
Here are some of the reference files:
output.avi is what we are looking for. best.onnx is the yolov8 model trained to detect the pantograph. cv_run.py file is running the object detection using opencv dnn module, a reference research paper is attached too describing an algorithm for finding the contact point between catenary and pantograph.
Please go through the files if you can and than let's discuss about costing and timeline.
The goal is to detect the correct contact point between pantograrph and catenary wire. Detecting the wires above pantograph should be first goal than finding the contact point between wire and panto and finally keep tracking this point.
Hope, you undersatnd the project clearly and think of a way to achieve it and come back to me with positive response.
Here is what I am expecting:
1. Find he contact point than keep detecting or tracking the correct contact point even switching of catenary wires in video input.
2. Making a full pipeline using yolo model to detect Pantograph, converting your model or algorithm to Onnx format and running inference with opencv dnn module using c++ code.
3. Testing on different videos or scenarios and overall accuracy which should be more than 85%. If needed reducing DL Model Size, optimization for Nvidia Board and obtaining more 15FPS speed.
Here are some of the reference files:
output.avi is what we are looking for. best.onnx is the yolov8 model trained to detect the pantograph. cv_run.py file is running the object detection using opencv dnn module, a reference research paper is attached too describing an algorithm for finding the contact point between catenary and pantograph.
Please go through the files if you can and than let's discuss about costing and timeline.
The goal is to detect the correct contact point between pantograrph and catenary wire. Detecting the wires above pantograph should be first goal than finding the contact point between wire and panto and finally keep tracking this point.
Hope, you undersatnd the project clearly and think of a way to achieve it and come back to me with positive response.