aruco_pkg

A repository for creating ArucoDetect, a set of models that assists ARCS in finding a physical marker and scanning it once close enough

Goals

The rover will be able to utilize a set of models that can detect and scan Aruco images, in conjunction with the ARCS, to detect when the rover has arrived at a destination.
ArucoDetect will determine the relative position of an Aruco marker to assist the ARCS with getting closer to the marker.

These are the strategies and technologies we will use to achieve the above goals

ArucoDetect is a set of CV models that will be used in the autonomous navigation challenge utilizing physical ARUCO markers.
We will research, develop, and deploy a set of CV models, pretrained and/or built and trained in-house, to the rover’s NVIDIA Jetson TX2 as a part of the ARCS.
ArucoDetect will interface with other portions of the ARCS to provide additional information about the position of a physical marker.
ArucoDetect will do this through utilizing CV models to analyze frames from the rover’s camera to detect a physical marker and determine the relative angle between the rover’s heading and the position of the marker.
ARCS will update its route as ArucoDetect updates the heading difference as the rover approaches the marker.
Once in range, ArucoDetect will scan the marker to indicate to ARCS that the rover has arrived at the position.

These are the products we will create as a result of achieving these goals.

A set of models that detects, finds the relative heading of, and scans Aruco markers, called ArucoDetect.
These models will be able to find the heading of the marker relative to the rover’s current heading to assist the ARCS in navigating to a position where the marker can be scanned, and the rover has arrived at the destination.