Motion Object Detection Algorithm for Sudden Illumination Changes

As the result of motion object detection can be greatly affected by sudden illumination changes,a background model based on local texture feature described by local binary pattern（LBP） which is gray-insensitive is proposed in this paper.Every coming frame and its background is converted into LBP first.The probability of a pixel belonging to foreground is generated by the operation of exclusive OR of the current image and the background’s LBP.Then the background can be updated adaptively and quickly based on the probability.Finally,we can get the foreground objects easily by a background subtraction.Experimental results demonstrated the effectiveness of the proposed approach in dealing with sudden illumination changes.It can relative precisely obtain moving objects and update background model quickly.