The use of “machine vision” technologies, including methods of signal processing theory and pattern recognition, are an alternative to GPS technologies (and in some cases, their complement) in the formation of the appearance of onboard high-precision integrated navigation and guidance systems for autonomous maneuvering of helicopter-type aircraft, such as multicopters. The practical applicability of “machine vision” technologies is largely based on advances in microelectronics and the development of sensory onboard technical devices in the visible, infrared and radar ranges of the electromagnetic spectrum, as well as on the possibility of implementing complex algorithms for the integrated processing of current information from several sources. The article considers various methods of image recognition as tools of “machine vision” technology, including the recognition methods based on the application of wavelet and Fourier transforms and their additive convolution. The choice of the optimal recognition method is justified by comparing them based on families of regular signals artificially created by shifting regarding to the base signal formed after extracting recognition features, image scaling and linearizing a given image of the Earth’s surface.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Application of Machine Vision Technology for Multicopter Flight Control Under Active Operation of Counter-UAV Systems

  • Ramin Rzayev,
  • Azer Kerimov,
  • Vagif Aliyev

摘要

The use of “machine vision” technologies, including methods of signal processing theory and pattern recognition, are an alternative to GPS technologies (and in some cases, their complement) in the formation of the appearance of onboard high-precision integrated navigation and guidance systems for autonomous maneuvering of helicopter-type aircraft, such as multicopters. The practical applicability of “machine vision” technologies is largely based on advances in microelectronics and the development of sensory onboard technical devices in the visible, infrared and radar ranges of the electromagnetic spectrum, as well as on the possibility of implementing complex algorithms for the integrated processing of current information from several sources. The article considers various methods of image recognition as tools of “machine vision” technology, including the recognition methods based on the application of wavelet and Fourier transforms and their additive convolution. The choice of the optimal recognition method is justified by comparing them based on families of regular signals artificially created by shifting regarding to the base signal formed after extracting recognition features, image scaling and linearizing a given image of the Earth’s surface.