基于深度摄像头的障碍物检测，代码的核心思路是首先通过二值化，将一米之外的安全距离置零不考虑，然后通过开运算去除掉一些噪点（这个后来发现不一定有必要），在求出所有障碍物的凸包，这个时候要计算面积，当面积小于一定的阈值的时候不予考虑，最终输出障碍物的凸包坐标(The core idea of the code is to first take the safe distance beyond one meter into account by binary, then remove some noise points by open operation(this later found to be not necessarily necessary), and find all obstacles in the convex packet. At this time, we must calculate the area. When the area is less than a certain threshold, the convex packet coordinates of the obstacle are eventually output)