MODEL OF AUTOMATED PROCESSING OF AIR RECONNAISSANCE DATA IN THE CONDITIONS OF APPLICATION OF UNMANNED AVIATION SYSTEMS

Possible ways of increasing the autonomy of unmanned aerial vehicles in the conditions of use for reconnais-sance purposes are being investigated. The possibility of integrating advanced information technologies into theformation and processing of aerial reconnaissance data generated by onboard optoelectronic systems of UAVs isbeing analyzed. Requirements are being formed for the automated processing of aerial reconnaissance data in theconditions of unmanned aircraft systems from the standpoint of ensuring the required level of reliability of intelli-gence information and the efficiency of its processing.An automated aerial reconnaissance data processing model is being developed for further integration onboardUAVs. The essence of this approach is to use open web platforms (Roboflow, Google Colab, Ultralytics HUB) forthe model’s preparation, deployment, and training. A distinctive feature of the proposed approach is the use ofcloud computing resources of the Google Colab web platform for training the model of automated detection of ob-jects of interest, which allows the obtaining of quantitative metrics for assessing the effectiveness of the developedmodel quite quickly.The use of the proposed method of forming an automated model allows for creating conditions for increasingthe autonomy of reconnaissance unmanned aerial vehicles by automating the process of processing aerial recon-naissance data in real-time; unified use of the trained model on board UAVs that perform typical tasks, i.e., to de-tect and recognize characteristic objects of interest; transformation of the model by the interests of aerial recon-naissance without the need for professional skills in the field of information technology.