Substantiation of the choice of the optimal UAV flight altitude for monitoring technological stresses for crops of winter rape

Abstract

Crop monitoring using UAVs is becoming standard practice in scientific research due to the possibility of an operational examination of large fields with high data quality due to high-resolution images. The need for operational inspection of crops on an industrial scale requires a certain balance between the permissible image quality and the acceptable monitoring time due primarily to the number of UAVs and their energy. The newest and most relevant stress factor for Ukraine is technological stresses. The aim of the study is the optimal flight altitude for monitoring stresses of a technological nature. Field studies were carried out from September to November 2019 in the Boryspil region on industrial crops of winter rape. Monitoring was carried out using the Slantrange 3p complex. Stresses were recorded according to 2 parameters — the area of the dome and the abnormal coloration of the 2 lower leaves of red, yellow, and a mixture of these colors. The maximum quality of identification is achieved at the maximum resolution of the images, that is, the minimum height, however, the issues of UAV energy determine the maximum height and, accordingly, the flight speed for the UAV. Thus, the optimal compromise is between the determination accuracy and the production-appropriate flight parameters, that is, a lot of metric optimization is needed. As a generalized criterion, the criterion of consistent products was chosen, where the number of local optimization criteria is s = 3, and the weighting coefficients of the criteria are selected subjectively by experts. The proposed algorithm for choosing the optimal UAV flight altitude to identify the technological nature of the stress of winter rape crops will allow planning monitoring activities and justifying the choice of UAV parameters relative to the duration of its flight. The type of objective function used will allow specialists without specialized education in programming to solve the problem of multi-criteria optimization by its own weight coefficients. The proposed approach has no restrictions on the number of optimization parameters and the number of local optimality criteria and can be used when the user chooses his own indication indicators.