• home
• current students
• graduate
• events

Presenter: Tunai Porto Marques
Supervisor:

Date: Fri, December 3, 2021
Time: 10:00:00 - 00:00:00
Place: ZOOM - Please see below.

ABSTRACT

Zoom meeting link: https://uvic.zoom.us/j/86127862723?pwd=RkxIa1M2SktncEVla3N1dnltekhNdz09

Meeting ID: 861 2786 2723
Password: 053496
Note: Please log in to Zoom via SSO and your UVic Netlink ID

Summary: Marine vessel traffic is one of the main sources of negative anthropogenic impact on marine environments. The automatic and efficient identification of boats in Environmental Monitoring (EM) images facilitates conservation, research and patrolling efforts. However, the various appearances of vessels, the highly dynamic nature of the water surface and weather-related visibility issues significantly hinder this task. While recent off-the-shelf Deep Learning (DL)-based object detectors can efficiently identify medium- and large-sized boats, smaller vessels, often responsible for substantial disturbance to sensitive marine life, are typically not detected. In this presentation I will describe a detection approach that combines state-of-the-art object detectors and a novel Detector of Small Marine Vessels (DSMV) to identify boats of diverse sizes, colors, and levels of visibility. The proposed DSMV uses a short time series of EM images and a novel bi-directional Gaussian Mixture Model (GMM) technique to determine motion in combination with context-based filtering and a DL-based image classifier. Experimental results obtained using our datasets of EM images containing boats of various sizes show that the proposed approach comfortably outperforms five popular end-to-end state-of-the-art object detectors. Code and datasets for this project are available at https://github.com/tunai/hybrid-boat-detection.