image rectification

[andrea-pilzer]

Hi, I am using the data_test.py script you provided in utils to rectify the images of the data split for scene parsing, because I want to train an unsupervised depth estimation model. I found there is a few pixel misalignment when trying to put side by side the two images (see images at the bottom).

I have a few questions:

• is it correct to use this code and its parameters for the scene parsing data split?
• can you suggest me how to adapt the script to rectify correctly the images?

Thank you.

The images I used (randomly picked from the dataset) are road02_ins/ColorImage/Record001/Camera 5/170927_063819921_Camera_5.jpg and road02_ins/ColorImage/Record001/Camera 6/170927_063819921_Camera_6.jpg This is zoom in the area where I found the problem: This are the two rectified images:

[XinJCheng]

We have updated the extrinsic params for stereo, shown as below: R = np.array([ [ 0.996797, 0.0384542, -0.0701246], [ -0.038475, 0.999259, 0.00105552 ], [ 0.0701132, 0.0016459, 0.997538 ]])

T = np.array([-0.636491,   -0.0158574,  -0.0599776])

Besides, we have tested the same picture with the new params, the results are shown as follow:

Detailed results:

Some comments: In order to get even better alignment results, you can downsample twice and our original image resolution is really high.

[andrea-pilzer]

Hi, thanks a lot for your answer! Now the sequences road02_ins and road03_ins are correctly rectified and I can use them with my model. Unfortunately I found that for sequence road01_ins these parameters are not correct.

[HobertXu]

I am working on the stereo-based depth estimation. I am confused about how to preprocess the data and rectify the image. In my way, the disparity doesn’t match with the corresponding camera-5 and camera-6 image pairs. Could you please give me some detailed suggestions about how to use this dataset?

[MJITG]

@HobertXu Hi， have you solved the problem? Need the stereo images to be rectified? I found them distorted very slightly...