Advanced Research Foundation: a competition for the best intellectual technology for decoding aerospace information
The development of intelligent data recognition systems is becoming increasingly relevant as the Internet of things develops and the rapid increase in the amount of information that devices collect and exchange: from security cameras to satellites surveying the Earth’s surface. However, the analysis of this information and the subsequent decision making remains for the man: due to natural limitations, he cannot quickly process large amounts of data and therefore remains a “bottleneck” in the process of collecting and managing information.
The rapid development of technical monitoring tools leads to a rapid increase in data volumes.
One of the key trends is to increase the information capabilities of the device sensors (including their sensitivity and radiometric accuracy): an unmanned aircraft generates gigabytes of data within 15 minutes of flight, and the amount of information collected by some meteorological gauges exceeds 100 GB per day.
')
Existing automated tools for thematic information processing use both newly arrived information (a specific image to be recognized) and relevant data accumulated, for example, in the course of previous observations. As a rule, the collection of visual data and their initial processing are carried out automatically, but then the obtained information is still analyzed by the human operator.
Intellectual information processing tasks assume that when solving them, a person will use his abilities to recognize the threats that have been formed during his evolution. In order to automate these processes, it is necessary to classify tasks that so far only a person can solve, to describe and characterize the capabilities of a human operator as fully as possible when solving them, and then draw up appropriate requirements for automatic thematic information processing. Today, intelligent visual processing technologies are being actively developed, for example, coming from CCTV cameras, as the human resource is not enough to process large amounts of data.
Development teams and teams of enthusiasts can try their hand at creating intelligent technology for the interpretation of specific information in the framework of the competition for the development of intellectual technology for image recognition and classification , which is currently being conducted by the Advanced Research Foundation and the Skolkovo Foundation. In this case, intellectual technology should be understood as an information system that can completely replace a human operator in the process of information processing.
The main objective of the competition is the development of technologies for automatic search and recognition of objects in images obtained during aerial photography and shooting from spacecraft in the visible wavelength range, that is, those that require the operator to fully analyze in current conditions. Today, the process of thematic interpretation of aerospace information consists of the phased grouping and transformation of data in order to create a specific picture of the earth's surface, focused on a specific task. Contestants will have to solve the problem of detecting objects in images, and images for analysis can be obtained both with aerial photography in the infrared range and with the use of Earth remote sensing spacecraft. The object of recognition will be considered a single rectangular structure, located outside human settlements. The structure of a single building and its location in relation to other buildings is determined by the expert marking of the training set, which will be provided to the participants by the organizers.
According to the results of the competition, the best participants will be able to become residents of Skolkovo without additional examination of their projects. The winner will also conclude an agreement on the implementation of the scientific and technical project of the Advanced Research Foundation for the creation of technology for automatic decoding of aerospace information. In addition, he will be offered funding from the Skolkovo Foundation in the amount of up to 5 million rubles.
The competition documentation for the object recognition problem on images obtained during aerial photography in the infrared range will be posted on the website of the Advanced Research Fund in the near future.
Write, participate, ask questions in the comments!
The rapid development of technical monitoring tools leads to a rapid increase in data volumes.
One of the key trends is to increase the information capabilities of the device sensors (including their sensitivity and radiometric accuracy): an unmanned aircraft generates gigabytes of data within 15 minutes of flight, and the amount of information collected by some meteorological gauges exceeds 100 GB per day.
')
Existing automated tools for thematic information processing use both newly arrived information (a specific image to be recognized) and relevant data accumulated, for example, in the course of previous observations. As a rule, the collection of visual data and their initial processing are carried out automatically, but then the obtained information is still analyzed by the human operator.
Intellectual information processing tasks assume that when solving them, a person will use his abilities to recognize the threats that have been formed during his evolution. In order to automate these processes, it is necessary to classify tasks that so far only a person can solve, to describe and characterize the capabilities of a human operator as fully as possible when solving them, and then draw up appropriate requirements for automatic thematic information processing. Today, intelligent visual processing technologies are being actively developed, for example, coming from CCTV cameras, as the human resource is not enough to process large amounts of data.
Development teams and teams of enthusiasts can try their hand at creating intelligent technology for the interpretation of specific information in the framework of the competition for the development of intellectual technology for image recognition and classification , which is currently being conducted by the Advanced Research Foundation and the Skolkovo Foundation. In this case, intellectual technology should be understood as an information system that can completely replace a human operator in the process of information processing.
The main objective of the competition is the development of technologies for automatic search and recognition of objects in images obtained during aerial photography and shooting from spacecraft in the visible wavelength range, that is, those that require the operator to fully analyze in current conditions. Today, the process of thematic interpretation of aerospace information consists of the phased grouping and transformation of data in order to create a specific picture of the earth's surface, focused on a specific task. Contestants will have to solve the problem of detecting objects in images, and images for analysis can be obtained both with aerial photography in the infrared range and with the use of Earth remote sensing spacecraft. The object of recognition will be considered a single rectangular structure, located outside human settlements. The structure of a single building and its location in relation to other buildings is determined by the expert marking of the training set, which will be provided to the participants by the organizers.
According to the results of the competition, the best participants will be able to become residents of Skolkovo without additional examination of their projects. The winner will also conclude an agreement on the implementation of the scientific and technical project of the Advanced Research Foundation for the creation of technology for automatic decoding of aerospace information. In addition, he will be offered funding from the Skolkovo Foundation in the amount of up to 5 million rubles.
The competition documentation for the object recognition problem on images obtained during aerial photography in the infrared range will be posted on the website of the Advanced Research Fund in the near future.
Write, participate, ask questions in the comments!
Source: https://habr.com/ru/post/313828/
All Articles