Strike-alert: Towards real-time, high resoltion navigational software for whale avoidance
Over the past few years, it has been shown that collisions with ships have become one of the major threats for whales. In order to reduce whale-ship strikes, we have started to develop schemes for identifying areas where whales are likely to be present in order to produce maps updated in real time for ships. Our case study is set in the Mediterranean Sea and our goal is to gather all the data available to improve our knowledge on whale distribution using machine learning techniques. The wide variety of data sources (e.g. very high resolution sensors on-board satellites, acoustical measurements, satellite tagging, direct reports from commercial ships, and social media along with streaming earth observation data) and the use of real time and streaming data will allow the development of high precision, real time maps of the likelihood of whale encounters. Our work seeks to dramatically improve the marine spatial effort by moving beyond ecological/environmental models to harness the full array of data and machine learning techniques. The driving idea is not to just create models of where strikes are likely to be, but to develop high resolution maps of probability of whale encounters in real time using all available data sources.
https://www.anthropocean.org/publications/pdf/strike-alert-towards-real-time-high-resoltion-navigational-software-for-whale-avoidance/view
https://www.anthropocean.org/@@site-logo/AnthropOcean logo.png
Strike-alert: Towards real-time, high resoltion navigational software for whale avoidance
Over the past few years, it has been shown that collisions with ships have become one of the major threats for whales. In order to reduce whale-ship strikes, we have started to develop schemes for identifying areas where whales are likely to be present in order to produce maps updated in real time for ships. Our case study is set in the Mediterranean Sea and our goal is to gather all the data available to improve our knowledge on whale distribution using machine learning techniques. The wide variety of data sources (e.g. very high resolution sensors on-board satellites, acoustical measurements, satellite tagging, direct reports from commercial ships, and social media along with streaming earth observation data) and the use of real time and streaming data will allow the development of high precision, real time maps of the likelihood of whale encounters. Our work seeks to dramatically improve the marine spatial effort by moving beyond ecological/environmental models to harness the full array of data and machine learning techniques. The driving idea is not to just create models of where strikes are likely to be, but to develop high resolution maps of probability of whale encounters in real time using all available data sources.