如何使用OpenCV基于含异常点的给定点集拟合矩形？

我有一组二维点集（可视化图如下），希望基于这些点近似拟合或绘制矩形。直接取最小x、最小y作为左上角，最大x、最大y作为右下角的方法无法处理异常点，以下是测试用示例数据：

[(569, 800), (569, 272), (404, 801), (404, 274), (624, 801), (624, 273), (457, 800), (457, 277), (516, 800), (516, 273), (686, 800), (686, 272), (681, 801), (681, 271), (343, 274), (352, 801), (399, 801), (399, 274), (349, 273), (341, 795), (442, 801), (442, 274), (462, 800), (462, 277), (415, 274), (433, 801), (411, 801), (420, 274), (431, 274), (451, 800), (451, 274), (676, 791), (344, 279), (676, 279), (347, 801), (356, 275), (483, 800), (342, 790), (676, 801), (473, 277), (344, 319), (667, 324), (346, 364), (667, 364), (347, 405), (664, 411), (345, 447), (666, 452), (387, 801), (387, 276), (347, 490), (654, 496), (409, 275), (342, 627), (655, 621), (342, 620), (666, 539)]

内容的提问来源于stack exchange，提问作者Ali Waqas