利用哈希表（字典）存储像素点出现的次数，同时记录出现次数为1的像素点的颜色值即可。

示例代码：

import cv2
import numpy as np

def find_different_color(img):
    # 将BGR图像转为RGB图像
    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
    height, width, _ = img.shape

    # 初始化哈希表
    color_dict = dict()

    # 遍历每个像素点并记录出现次数
    for row in range(height):
        for col in range(width):
            color = tuple(img[row, col])
            if color in color_dict:
                color_dict[color] += 1
            else:
                color_dict[color] = 1

    # 遍历哈希表找出只出现过一次的像素点的颜色值
    color_list = []
    for color, count in color_dict.items():
        if count == 1:
            color_list.append(color)

    return color_list


if __name__ == '__main__':
    img = cv2.imread('test.png')
    color_list = find_different_color(img)
    print(color_list)

注：以上代码以Python语言和OpenCV库为例，步骤可以根据具体情况进行调整。