To segment a binary seal image out of a check image without distortion,an adaptive morphological segmentation algorithm based on a top-hat transformation is proposed to accurately extract a binary Chinese square seal from a bank check. A gray scale square seal is extracted from a coloured bank check according to the colour information. Different Chinese characters have different stroke features and background evenness. To respectively process each character in the square seal, the seal is divided into four sub-squares. Then,the background across each sub-square of the grey scale seal image is smoothed by the top-hat transformation. The size of the structuring element in the top-hat transformation can have a significant influence on the segmentation. So an adaptive method is proposed to iteratively estimate the proper size of the structuring element according to the local foreground area. In each sub-square, the optimal size of the structuring elements, respectively, for the imprint frame and the character are estimated. With their optimal structuring elements, the character and the imprint frame in each sub-square are filtered by the top-hat transform and binarized by the Otsu’s method. The experimental result shows that when 350 different square seals are segmented in bank checks,only 2 segmented seals have distortions.It is proved that the proposed algorithm can correctly segment Chinese characters with intricate and dense strokes in a bank check square seal. Adhesion and incompleteness distortions in the segmentation results are reduced, even when the original square seal has poor quality.