Vision-based object recognition under atmospheric distortions

MyWorld researchers in Bristol Vision Institute have been awarded £340,000 from the UK Defence and Security Accelerator programme to develop new methods for vision-based object recognition in the presence of atmospheric distortions.

Heat haze and atmospheric turbulence can significantly reduce the quality of visual imagery and degrade the effectiveness of conventional approaches to detection and classification which, in turn, makes scene interpretation and analysis extremely difficult. This is particularly important for natural history filmmaking as well as for security applications. This new two-year project led by Dr Pui Anantrasirichi, Professor David Bull and Professor Alin Achim, will address this challenge using modern data-driven machine learning methods. It will develop real-time video restoration techniques that mitigate spatio-temporal distortions and thus improve visual interpretation of the scene by a human observer. It will also assist decision making by implementing and evaluating real-time object recognition and tracking using the restored video.