The global shipping industry is responsible for moving billions of dollars worth of goods and raw materials around the planet every day. Closer to home smaller boats form the backbone of many cities' transportation systems. It is no surprise then that many AI companies are attempting to make water going vessels safer and more efficient with the help of computer vision AI.
AI systems can pilot boats autonomously. As a result AI powered boats should be cheaper to operate and capable of more individual journeys. However, there are still numerous challenges that autonomous ships need to overcome before they are widely adopted. Data annotation services, like Keymakr, can play an important role by providing developers with high quality training datasets.
This blog will first look at the potential uses for autonomous ships. Secondly we will address some of the complex challenges facing the technology. And finally we will show how data annotation providers can help.
The need for autonomous ships
Most ships currently in operation are pollution emitting and difficult to staff. AI equipped vessels could help the industry by reduced emissions and improving cost efficiency in a number of areas:
- Waste disposal: Some cities are so clogged with traffic that it can be hard for refuse to be transported. Water based waste disposal may be an answer for this. Autonomous boats could remove waste from city centres using canals and other waterways.
- Package delivery: Rivers and canals were often used in the past for mail and package delivery. AI could make this a possibility in the future. Fleets of autonomous boats could transport packages through the night as part of an integrated delivery system.
- Ferry services: Short, regular journeys are an important function of boats. AI could easily be used to automate simple ferry services. This might mean moving people across a river in a city, or operating as a tour boat for visitors.
- International shipping: The holy grail of autonomous ship development is a fully AI guided international shipping fleet. If this becomes a reality it could mean cheaper transportation prices globally.
Challenges to overcome
AI powered boats could change the way we navigate seas and waterways. However, developers still need to deal with a number of persistent problems:
- In water obstacles: Boats cannot stop as quickly as cars. This means they have to be particularly aware of their surroundings so that they can navigate around obstacles. There are a wide range of potential things that boats could hit: other boats, swimmers, branches, animals and large pieces of garbage. AI systems need to be capable of identifying and avoiding all of these obstacles.
- Bridges: Bridges can present a challenge in tight city waterways. Boats need to be able to navigate under their highest point and avoid crashing into lower sections.
- Challenging passages: For ships travelling longer distance coastal landscapes can be a challenge. Computer vision can help autonomous ships to avoid potential collisions with sandbanks and headlands.
How data annotation services can help
The promise of AI powered boats is clear. However, AI companies need the help of smart data annotation services like Keymakr to ensure they have access to the best training datasets.
- Varied datasets: Keymakr has expertise in data collection and creation. As a result we easily create varied datasets that help models learn effectively.
- Project management: Keymakr’s project management systems ensure that annotation tasks are given to the highest performing team members.
- Semantic segmentation: Keymakr specialises in adding to semantic segmentation to images and video.