Target real-time monitoring method for underwater unmanned aerial vehicle

Abstract

The invention belongs to the technical field of target real-time monitoring, and discloses a target real-time monitoring method for an underwater unmanned aerial vehicle. The method comprises the following steps: firstly, placing an unmanned aerial vehicle device in a monitoring target environment and performing initializing, converting the specific position of a monitoring target into data required to be monitored, performing probability density modeling on the data of the monitoring target, further updating the obtained model on deep learning model parameters of the underwater unmanned aerial vehicle; and then changing the position of the monitoring target and continuing the operation in the previous step until the whole monitoring environment parameters are completely integrated, comparing the collected data models, and if the comparison result does not meet the convergence condition, obtaining a Mean-Shift vector through calculation, and adjusting the monitoring data. According to the invention, general monitoring of water area environment water quality parameters and the like can be carried out in time, the problem of great change of monitored target data is solved, and the credibility and accuracy of the monitored data are improved.

Description

technical field [0001] The invention belongs to the technical field of real-time monitoring, in particular to a targeted real-time monitoring method of an underwater unmanned aerial vehicle. Background technique [0002] Targeted real-time monitoring of UAVs refers to placing underwater UAVs in a three-dimensional monitoring environment, sampling data for monitoring targets, and establishing the optimal solution as monitoring data by building a model. It has a good application prospect in the fields of water quality detection, real-time communication and human-computer interaction. [0003] Due to the detection of UAVs in water, the operation and real-time communication of UAVs are unstable, resulting in inaccurate monitoring data, etc., which poses great challenges to UAV monitoring. [0004] In related technologies, some scholars have proposed some end-to-end, unsupervised advanced monitoring algorithms and models by studying traditional correlation filtering algorithms a...

AI Technical Summary

Problems solved by technology

The disadvantage of related technologies is that these methods require high computing resources and are not suitable for real-time monitoring of UAVs.

[0005] In addition, during the monitoring process of the Mean-Shift algorithm in the related art, the core width of the template remains unchanged. When the limit coordinates of the monitored water area change due to underwater factors, the real core width of the new target may change.

Images