The way of deep learning thermodynamic diagram visualization

heatmap： It's the heat map you want to superimpose on the original image
img ： It's your original picture

1. First of all, will heatmap The maximum value of is reduced to 0-255 Between unit8 type
2. And then heatmap Convert to color image
   You can use from cv2 The function in

heatmap = cv2.applyColorMap(heatmap,cv2.COLORMAP_JET)

3. take heatmap and img weighting

superimposed_img = (heatmap*0.3+img*0.7).astype(np.uint8)

Yes superimposed_img You can visualize