ARTIFICIAL INTELLIGENCE AND ROBOTICS

Geomnibus Newsroom
May 29, 2020

Themen

Due to the extraordinarily rapid advancement of technology and computing power in recent years, artificial intelligence and robotics are more relevant and in demand today than ever before. The worldwide sales of software companies in the field of artificial intelligence are predicted to increase from US$5.4 billion in 2017 to around US$105.8 billion in 2025.

Based on the applications presented at INTERGEO 2018, artificial intelligence is mainly used to analyze imagery and big data, as well as in measuring instruments that generate and process massive amounts of highly complex data. Artificial intelligence is not only facilitating such processes, but also hugely accelerating them. In satellite image analysis, for example, machine learning is primarily used for automatic analysis and feature extraction. With over 300 satellites currently observing the Earth and continuously producing digital images, machine learning makes it possible to analyze such massive amounts of imagery on a large scale and in near real time. The key advantages over manual evaluation are therefore accuracy, speed and ease of use. However, the major challenges include procurement of a machine learning solution, as well as the quality and reliability of the data used to train the algorithms, since there is no uniform solution for this.

The current options are either to generate artificial training data or to purchase it from other providers. In addition, artificial intelligence is used in total stations for on-device surveying parameters. The device automatically records the measuring conditions and adjusts the necessary settings, such as light settings, ac-accordingly. In the case of sensors or measuring devices that often suffer from increased ‘noise’ due to weather conditions, artificial intelligence can help to eliminate or at least reduce measurement errors. In addition, sensor data such as traffic density can be predicted very accurately using suitable training data. In contrast to artificial intelligence, the use of robotics was limited mainly to the area of UAVs at this year’s INTERGEO. Since 2015, small marking and staking robots have been available to support surveyors and road builders. However, these are not autonomous but must be pre-program-med for certain tasks.

Rather than accuracy, their key benefits are that – unlike humans – robots do not become tired and they can be used more flexibly in harsh environments. In addition, robots can now be remotely controlled from a tablet or smartphone, which makes them easier to use.

Jetzt diese Artikel lesen