AP08856412 "Development of intelligent data processing and flight planning models for solving the problems of precision farming using UAVs"

Relevance

The use of UAVs for solving a wide range of monitoring and control tasks in the economic sectors of Kazakhstan is limited not only by the individual technical features of this mobile platform, but also by the insufficient development of practically applicable intelligent methods, algorithms and systems for traffic control and analysis of data coming from the UAV. The project is aimed at developing practically applicable methods that provide solving the problems of flight control (including a group of devices), identification and classification of objects of observation, using modern machine learning methods to solve precision farming problems. Expected results are applicable in other industries to solve monitoring problems.

Objective of the project

Development of data processing and flight planning models for technically heterogeneous UAVs for solving precision farming problems based on artificial intelligence methods.

The expected results of the project are

As a result of the project will:

A project website has been developed, thanks to which potential consumers and the scientific community will be able to get acquainted with the results of scientific research work.

The results of the project will have a high applied result and will be a significant contribution to the development of applied machine learning methods. The project will develop important elements of intelligent unmanned technology in terms of computer vision and flight planning.

Scientific results can be applied or commercialized as part of the final product in the tasks of management and monitoring in agriculture and other areas of UAV economic application.

The scientific and technical effect as a result of fulfilling the tasks outlined in the project is to develop elements of a new technology based on intelligent data processing and a higher degree of process automation. This effect can be multiplier, as the technologies being developed may be applied not only in agriculture, and the options for its use may require additional research and development.

Results achieved

Additional sets of video data were generated, high-quality video data and photo images were recorded from the UAV, synthetic and real data sets were obtained in the amount of several hours of video in different lighting conditions. Methods for preprocessing video and other data (photo images) received from UAVs have been developed, methods for preprocessing data for synthetic and real data have been obtained and tested, implemented in the form of software modules and suitable for solving problems of precision farming. The adaptation of existing machine learning methods for solving the problems of processing data received from an UAV on a ground-based computer complex was carried out. Methods have been developed for solving image recognition problems and detecting differences between frames using pre-trained models (YOLO v3 Darknet, Alexnet, Inception v3, ResNet 50). The ML methods (YOLO v3 darknet) were adapted for image processing on board the UAV. A software environment has been developed for conducting computational experiments and simulations of overflights of a given territory by a group of drones in order to perform the tasks of video / photography, spraying and search, an algorithm for optimizing the plan for flying over a territory by a group of drones has been implemented, taking into account a number of optimization criteria. To solve the optimization problem, a genetic algorithm was used. Flight planning, identification and classification systems were tested in a simplified environment, a test report was signed.

Names and surnames of the study group members with their identifiers

1. Mukhamediev Ravil Ilgizovich - scientific adviser
2. Kuchin Yan Igorevich - senior researcher
3. Yakunin Kirill Olegovich - senior researcher
4. Symagulov Adilkhan - engineer
5. Murzakhmetov Sanjar Bakhtierovich - software engineer
6. Bekbagambetov Abai - software engineer
7. Asanov Ilyas - junior research fellow
8. Ospanova Maryam - junior research fellow
9. Elis Marina - junior research fellow

List of publications

1. R. Muhamedyev, K. Yakunin, Y. Kuchin, A. Symagulov, S. Murzakhmetov, M. Ospanova, I. Assanov, M. Yelis Intelligent unmanned aerial vehicle technologies // The 18th Int. conf. Information technologies and management. -  Riga, 2020. - P.21-22.
2. Ravil Mukhamedyev, Yan Kuchin, Kirill Yakunin, Adilkhan Symagulov, Maryam Ospanova, Ilyas Assanov, Marina Yelis. Intelligent unmanned aerial vehicle technology in urban environments //Int. Conf. on Digital Transformation and Global Society. – Springer, Cham, 2020. –  16 p. ( Q3, Scopus Cite Score = 32%).
3. Mukhamediev RI, Symagulov A, Kuchin Y, Yakunin K, Yelis M. From Classical Machine Learning to Deep Neural Networks: A Simplified Scientometric Review//Applied Sciences. – 2021. – Т. 11. – №. 12. – P. 5541. (CiteScore highest quartile = Q2, JCR - Q2, CiteScore =3.0, CiteScore highest percentile=71%, WoS IF=2.679)
4. Mukhamediev R. I. et al. Review of Some Applications of Unmanned Aerial Vehicles Technology in the Resource-Rich Country //Applied Sciences. – 2021. – Т. 11. – №. 21. – С. 10171. https://doi.org/10.3390/app112110171 (CiteScore highest quartile = Q2, JCR - Q2, CiteScore =3.0, CiteScore highest percentile=71%, WoS IF=2.679)
5. Mukhamediev R. I. et al. Rapid bibliometric analysis in deep learning domain //2021 Int. Conf. on Information and Digital Technologies (IDT). – IEEE, 2021. – P. 206-211.
6. I. Assanov. Multi UAV simulator in Unity // The 19th Int. conf. Information technologies and management. -  Riga, 2021. - P.46-47.
7. A Bekbaganbetov, M Ospanova, M Yelis, J Rabcan, R Muhamedyev. Experiments to identify changes in synthesized images // The 19th Int. conf. Information technologies and management. -  Riga, 2021. - P.54-55.
8. M. Ospanova, M. Yelis, A. Bekbaganbetov, J. Rabcan, R. Muhamedyev Image generation for solving problems of precision farming // The 19th Int. conf. Information technologies and management. -  Riga, 2021. - P. 64-65.
9. P Sedlacek, M Ospanova, M Yelis. Sensitivity analysis of MVL Systems by the Logic Derivatives of MVL Functions //CERes Journal. - 2020.- Vol.6, № 2.- P.1-7.