APPLICATION OF MULTI-KERNEL APPROACH INBACKGROUND SUBTRACTION METHOD UNDER DYNAMICILLUMINATION CONDITIONS

This study addresses the application of different kernel sizes and shapes to optimize morphological operations in the process of background subtraction in video.In this approach, a multi-kernel strategy is employed to adapt to various environmental conditions, such as illumination and object sizes.In this way, the accuracy of object detection is enhanced by adaptively adjusting the morphological operations using a multi-kernel strategy to refine the results of background subtraction.