TRENDS AT INTERGEO
The trend towards cloud-based storage and processing of increasing volumes of data continues unabated. Many cloud providers’ services now go beyond data storage to offer customers the “full package” of processing and analysing, even visualising and releasing, data on their behalf. One case in question is saving large point clouds on a cloud with a user-friendly interface, enabling large volumes of data to be uploaded and then viewed or processed there. Exhibitors cited the major advantages of this system as doing away with the need for separate software, or even to install hardware, and that the data can be downloaded in an up-to-date format at any time; not to mention that it’s easier for non-specialists to call up and visualise data in this manner.
While these “full packages” offered by cloud providers are certainly very practical, in the same way as IoT data, they also give rise to serious reservations in relation to data protection and the security of data stored this way. Since the General Data Protection Regulation came into force in 2018, many companies and cloud providers have this year made a point of the fact that they store cloud data exclusively on servers located and maintained in Germany or the European Union. Artificial intelligence (AI), particularly machine learning, was generally used less intensively in the past in GIS applications than in remote sensing or photogrammetry. By now, it has started cropping up all over the place.
Besides the usual applications such as processing large volumes of data in the cloud, AI is used to help detect damage to roads, for example (See: “Mobile mapping”). At INTERGEO, solutions incorporating AI concepts were also exhibited for predicting specifically located events (such as detecting mineral deposits), forest pests, erosion, or natural hazards. Local authorities also use AI for chat bots to receive residents’ messages via smartphone and swiftly offer solutions or put them in touch with the relevant departments. Despite their numerous benefits, AI concepts often require human interaction in many cases to minimise errors and for quality assurance regarding the outcome. For example, in the semi-automatic AI concepts used to extract vector objects from raw data, algorithms merely assist the humans who check the results and possibly improve them. In solutions that do not require human intervention, the algorithms only handle a limited scope of tasks. Apart from that, it was notable that businesses were generally far more open than local authorities to applications involving of AI methods.11.