snapshot...

modules/deeplearning/cnn_cld_frac.py

@@ -175,27 +175,47 @@ def upsample(tmp):
     return tmp
 
 
+def upsample_mean(grd):
+    bsize, ylen, xlen = grd.shape
+    grd = get_grid_cell_mean(grd)
+    up = np.zeros((bsize, ylen, xlen))
+
+    up[:, ::2, ::2] = grd[:, :, :]
+    up[:, 1::2, ::2] = grd[:, :, :]
+    up[:, ::2, 1::2] = grd[:, :, :]
+    up[:, 1::2, 1::2] = grd[:, :, :]
+
+    return up
+
+
+def get_grid_cell_mean(grd_k):
+    grd_k = np.where(np.isnan(grd_k), 0, grd_k)
+
+    a = grd_k[:, 0::2, 0::2]
+    b = grd_k[:, 1::2, 0::2]
+    c = grd_k[:, 0::2, 1::2]
+    d = grd_k[:, 1::2, 1::2]
+    s = a + b + c + d
+    s /= 4.0
+
+    return s
+
+
 def get_label_data(grd_k):
     grd_k = np.where(np.isnan(grd_k), 0, grd_k)
-    # grd_k = np.where(grd_k < 0.5, 0, 1)
 
     a = grd_k[:, 0::2, 0::2]
     b = grd_k[:, 1::2, 0::2]
     c = grd_k[:, 0::2, 1::2]
     d = grd_k[:, 1::2, 1::2]
     s_t = a + b + c + d
-    # s_t = np.where(s_t == 0, 0, s_t)
-    # s_t = np.where(s_t == 1, 1, s_t)
-    # s_t = np.where(s_t == 2, 1, s_t)
-    # s_t = np.where(s_t == 3, 1, s_t)
-    # s_t = np.where(s_t == 4, 2, s_t)
     s_t /= 4.0
 
     blen, ylen, xlen = s_t.shape
     s_t = s_t.flatten()
     cat_0 = np.logical_and(s_t >= 0.0, s_t < 0.2)
-    cat_1 = np.logical_and(s_t >= 0.2, s_t < 0.8)
-    cat_2 = np.logical_and(s_t >= 0.8, s_t <= 1.0)
+    cat_1 = np.logical_and(s_t >= 0.2, s_t < 0.7)
+    cat_2 = np.logical_and(s_t >= 0.7, s_t <= 1.0)
 
     s_t[cat_0] = 0
     s_t[cat_1] = 1