在点周围创建方形缓冲区 - 网格数据 - python

Question

我有一个数据集，我想为其创建额外的训练标签，方法是在二维数据集（经度、纬度）中的真实标签周围创建一个缓冲区。为了我的问题，假设我的数据集看起来像：

array([[0, 0, 0, 0, 0],
       [0, 0, 0, 0, 0],
       [0, 0, 1, 0, 0],
       [0, 0, 0, 0, 0],
       [0, 0, 0, 0, 0]])

代码：df = np.array([0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0]).reshape(5,5)

创建缓冲区后。我的输出数据应该类似于：

array([[0, 0, 0, 0, 0],
       [0, 1, 1, 1, 0],
       [0, 1, 1, 1, 0],
       [0, 1, 1, 1, 0],
       [0, 0, 0, 0, 0]])

从技术上讲，我的数据集是具有 5000 个时间变量的 3D 数据集。我知道在 ArcGIS 中有一个工具可以做到这一点。但是，它一次只执行一次。正如你所理解的，我不想导出 5000 个单独的文件。有谁知道如何解决这个问题？

也许很高兴知道我所有的一个“像素”都是 0.5 x 0.5。

Answer 1

虽然它可能不是最漂亮的答案。我确实找到了解决方法。下面的代码（如果可能）在原始数据集中的每个真实标签周围创建一个 3x3 真实标签网格。它也可以毫无问题地处理边界和边缘。如果有人知道更好/更快的解决方案，请分享！

df = np.array([0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0]).reshape(5,5)

values = np.where(df == 1) # find indexes of true labels
    for ind in range(len(values[0])):
        x = values[0][ind]
        y = values[1][ind]
        #upper left
        if x == 0 and y == 0:
            df[x,y:y+2] = 1
            df[x+1,y:y+2] = 1 
        #upper right
        elif x == 0 and y == df.shape[1]-1:
            df[x,y-1:y+1] = 1
            df[x+1,y-1:y+1] = 1
        #bottom left
        elif x == df.shape[0]-1 and y == 0:
            df[x-1,y:y+2] = 1
            df[x,y:y+2] = 1
        #bottom right
        elif x == df.shape[0]-1 and y == df.shape[1]-1:
            df[x-1,y-1:y+1] = 1
            df[x,y-1:y+1] = 1
        ### along borders
        #along top border
        elif x == 0 and y < df.shape[1]-1:
            df[x,y-1:y+2] = 1
            df[x+1,y-1:y+2] = 1
        #along bottom border
        elif x == df.shape[0]-1 and y < df.shape[1]-1:
            df[x-1,y-1:y+2] = 1
            df[x,y-1:y+2] = 1
        #along left border
        elif x < df.shape[0]-1 and y == 0:
            df[x-1,y:y+2] = 1  
            df[x,y:y+2] = 1
            df[x+1,y:y+2] = 1
        #along right border
        elif x < df.shape[0]-1 and y == df.shape[0]-1:
            df[x-1,y-1:y+1] = 1  
            df[x,y-1:y+1] = 1
            df[x+1,y-1:y+1] = 1
        ### everywhere aside along borders
        else:
            df[x-1,y-1:y+2] = 1
            df[x,y-1:y+2] = 1
            df[x+1,y-1:y+2] = 1