Extraction of the objects observed on a non-uniform background during sensor motion

The work is addressed to the problem of an object extraction in the images during sensor motion. In the case the object extraction requires the registration of the observed image and the reference background image. The error of the registration causes false alarms in the extraction result. In this paper the problem of object extraction during sensor motion is solved by taking into the consideration the statistical properties of this error. The solution of this problem has been obtained using the method based on Johnson distribution parameters estimation. This method has very high computational complexity; therefore, the simplified algorithm for an object extraction was developed. The experimental research results are also presented.