Shulou(Shulou.com)11/24 Report--

Thanks to CTOnews.com netizens Xiao Zhan cut, West window old clues delivery! CTOnews.com, October 20, according to an article posted by DAMO official account, Ali Dama today released the industry's first remote sensing AI large model (AIE-SEG), which claims to be "the first to achieve the task of image segmentation in the field of remote sensing", "a model to achieve the rapid extraction of 'zero samples of all things', and can identify nearly 100 kinds of remote sensing features such as farmland, waters, buildings and so on. The recognition result can be adjusted automatically according to the interactive feedback of the user.

It is reported that remote sensing technology is mainly used in urban planning, cultivated land protection, emergency relief and other industries, while with the blessing of AI, relevant remote sensing technology can analyze satellite capture content and historical meteorological data, thus helping urban operation, cultivated land protection, emergency disaster relief and other industry applications.

CTOnews.com summarizes the characteristics of the large remote sensing model as follows:

Support multimodal interaction, such as entering "extract cultivated land from image", the selected target will be automatically identified.

▲ Map Source Dama DAMO official account supports arbitrary surface target recognition and establishes a multi-level semantic tagging system

▲ Image Source Dharma Hall DAMO official account supports full element extraction including satellite and UAV images

▲ image source DAMO official account supports interactive result correction. If the "water area" is also extracted when identifying "paddy field", the deviation can be corrected by manual operation.

▲ Tu Yuan Dama DAMO official account supports general and multi-category change detection

The official DAMO official account of ▲ Image Source Dama said that in some specific scenes, compared with the traditional remote sensing model, the accuracy of case extraction can be improved by 25%, and the accuracy of change detection can be improved by 30%.

The Dharma Academy also claimed that the large remote sensing AI model provides "out-of-the-box" API calling service, and users can customize different remote sensing AI interpretation functions according to their needs, such as "water extraction", "cultivated land change monitoring", "photovoltaic identification" and so on.

This will allow AI to sink further into the field and greatly improve the analysis efficiency of remote sensing applications such as disaster prevention, natural resource management, agricultural yield estimation and so on.

At present, the AI model has been applied in the industry. for example, Shandong Land surveying and Mapping Institute and Aridamo Institute have cooperated with remote sensing AI model to monitor the growth of winter wheat, with an identification accuracy of more than 90%, which can effectively improve the efficiency of remote sensing interpretation of winter wheat and help agricultural managers better predict grain yield and improve agricultural production efficiency.

The National Institute of Natural disaster Prevention and Control also uses this model to identify landslides and collapsed buildings. in the test of remote sensing images of historical natural disasters, it only takes more than ten minutes to extract disaster information, which is tens of times more efficient than manual identification. to provide efficient and accurate remote sensing analysis support for scientific disaster relief.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.