mukhlasadib / kitti_mapping Goto Github PK

Hi, thanks for the detailed explanation of the code. I would like to know how the pretrained model has been trained for road and non-road classes. Is it manually annotated or available from KITTI open source?

Description

I encountered an IndexError while running a road segmentation function using the DeepLab v3+ model on an image. The function is designed to process camera images for road segmentation and map LiDAR points to the segmented image to identify road points. However, when executing the code, it throws an IndexError indicating an invalid index to a scalar variable.

Error Details

The error traceback is as follows:

---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
Cell In[21], line 52
     49 sess = tf.compat.v1.Session(graph=graph)
     51 DEEPLAB_INPUT_SIZE = 513
---> 52 segm_prob, segm_pred = process_images(crop_img, sess, DEEPLAB_INPUT_SIZE, 0.5)
     53 segm_prob1, segm_pred1 = process_images(crop_img1, sess, DEEPLAB_INPUT_SIZE, 0.5)
     55 ### Visualize

Cell In[21], line 37
     35 segm_reg[segm_reg==0] = np.nan
     36 modes,_ = stats.mode(segm_reg.flatten(), axis=None, nan_policy="omit")
---> 37 mode = modes[0]
     38 pred[segm_reg!=mode] = 0
     40 return prob, (pred*255).astype(np.uint8)

I want to know how to solve this problem, please help me, thank you!

Would you please provide some explanation of how this shift_pose_dgm() method work? Thanks in advance.