gui miss matching

[wuhanhui]

I check the code slam_interface_gui.py found the gui is not like image above such as closure l. Isn’t it that the code is not up to date? And what type of image does this depth_maps point to? Can it point to the code generated by batch_processing.py?

[GuidoManni]

Hi @wuhanhui!

Thanks for reporting this issue.

Regarding the GUI interface: the images in the documentation show an older version of the interface. The current slam_interface_gui.py implementation has been simplified and those specific UI elements are no longer present in the codebase.

Regarding depth maps from batch_processing.py:
Yes, the depth maps generated by batch_processing.py are fully compatible with the SLAM interface. The batch processing script outputs PNG depth files that can be used directly as the depth_path parameter, as long as you have corresponding RGB images with matching filenames.

The system expects:

• Depth files in PNG format (which batch_processing.py provides)

• Sequential naming that matches your RGB files

• Equal numbers of RGB and depth images

So you can definitely use the output from examples/depth_estimation/batch_processing.py for your 3D reconstruction workflow.

[wuhanhui]

Thanks for your reply. I found the images generated by batch_processing in the output folder is Shape: {img.shape} is : Shape: (480, 600, 4), But min_val, max_val, _, _ = cv2.minMaxLoc(depth_image) from slam_utils.py surppose the depth_image is a single channel image. So I meet the error

OpenCV(4.12.0) D:\a\opencv-pytsrc\minmax.cpp:1121: error: (-215:Assertion failed) (cn == src\minmax.cpp:1121: error: (-215:Assertion failed) (cn == 1 &src\minmax.cpp:1121: error: (-215:Assertion failed) (cn == 1 && (_mask.empty() || _mask.type() = && (_mask.empty() || _mas _mask.empty() && !minIdx && !maxIdx) in functio _mask.empty() && !minIdx && !maxIdx) in functio _mask.empty() && !minIdx && !maxIdx) in function 'cv::minMaxIdx'
_mask.empty() && !minIdx && !maxIdx) in function 'cv::minMaxIdx'

I am using opencv 4.12 and python 3.8