Machine vision systems for detection of fast-moving objects in low visibility conditions

Machine vision technology for the detection of fast moving objects in low visibility conditions requires a more careful approach to the optical system, image processing tools, cameras based on a silicon and indium gallium arsenide focal plane matrix in the infrared range, which take into account the magnitude and variation of atmospheric brightness, ground illumination and evaluation of sensitivity and cameras in the terahertz range. Since the main focus is on aerial objects that have low visibility, especially in the dark, the use of infrared cameras has become the standard. The movement of the object, the shortcomings of the optical system create additional difficulties when processing the effects of interference, thermal noise of cameras, respectively, the volume and time of calculations increases, which play a key role in real-time systems for detecting and tracking moving targets. In order to take into account the spectral characteristics of cameras and the influence of external factors, neural networks with deep learning are applied with the maximum use of image processing packages.