|
An approach to the thermo vision image preprocessing
A new approach to primary thermal image processing method has been proposed in this paper. The object of the preprocessing is the image generated by the first generation thermal image camera that uses scanning by vertical array of detectors. This type of image contains a certain number of artificial lines, known in the references as ‘false lines’. These inserted lines do not contain any useful information, and may cause some additional problems in the process of target detection and tracking. The proposed method is based on the analysis of the image auto-correlation function. This function is used as indicator of false lines existing, FIR (Finite Impulse Response) filters are used for image preprocessing in order to decrease the influence of false lines, and as a quality of the algorithm the smoothness of the auto-correlation function has been considered. The drawback of the method is decreasing of the image contrast. This contrast is smaller as the order of the filter is higher. High contrast is required in order to detect and track the targets efficiently. These two contradictory effects call for a compromise choosing the filter order. The results of the proposed algorithm have been illustrated through the processing of the real thermo vision image.
Key words: thermovision, image processing, image creation, image quality, target detection, target tracking.
|