An Embedded Monocular Vision Approach for Ground-Aware Objects Detection and Position Estimation

TL;DR

This paper presents an embedded monocular vision system using SSD MobileNet v2 on NVIDIA Jetson Nano for real-time detection and position estimation of soccer objects, outperforming existing systems within 1 meter range.

Contribution

It introduces a ground-aware, monocular vision approach optimized for embedded systems, enabling accurate, real-time soccer object detection and localization.

Findings

Achieves 30 fps processing speed.

Overcomes existing SSL vision systems for objects within 1 meter.

Root Mean Square Error of 14.37 mm for ball localization.

Abstract

In the RoboCup Small Size League (SSL), teams are encouraged to propose solutions for executing basic soccer tasks inside the SSL field using only embedded sensing information. Thus, this work proposes an embedded monocular vision approach for detecting objects and estimating relative positions inside the soccer field. Prior knowledge from the environment is exploited by assuming objects lay on the ground, and the onboard camera has its position fixed on the robot. We implemented the proposed method on an NVIDIA Jetson Nano and employed SSD MobileNet v2 for 2D Object Detection with TensorRT optimization, detecting balls, robots, and goals with distances up to 3.5 meters. Ball localization evaluation shows that the proposed solution overcomes the currently used SSL vision system for positions closer than 1 meter to the onboard camera with a Root Mean Square Error of 14.37 millimeters. In addition, the proposed method achieves real-time performance with an average processing speed of 30 frames per second.