Skip to content

PoliTOcean/Stereo-distance-estimation

Repository files navigation

Stereo-distance-estimation

Interactive interface to estimate length using stereo camera and SuperGlue model.

Setup the Environment

Create Python environment

python -m venv env

Activate the environment

source env/bin/activate

Install requirements

pip install -r requirmenets.txt

Instruction Manual

Run the script

python super_glue.py

For the first part you need to select between 2 and 4 points on the image using the mouse leftclick, if you want to deselect your last point use mouse rightclick. When you are done with the selection press ESC to go to the next stage. You need at least 2 points, if not enough points are selected you need to do this phase again.

Then the script will match your selected points with the keypoints on the image and it will compute the disparity. If a match is not found or the computation of the disparity fails, you need to select your points again.

The third step consists of chosing if the distance is computed using the single point or the middlepoint of twos. You need to make 2 choices: press 's' if you want to use the first (or third) point as first edge or press 'm' if you want to use the middle point of the first and second selection (or the third and forth selection).

After this step the script will compute the lenght distance between the two edge points, the value is printed on the terminal.

At every step information are available on the terminal.

The script

distance_manual.py

let's manuallt select all the point from left and right images: first select a point on the left image, then select the same point on the right image manually and so on, it will compute the disparity and the lenght of the segment.

use

stereo_calibration.py

to compute the intrinsic and extrinsic parameters of the stereo camera that are then used by the other scripts.

About

Interactive interface to estimate length using stereo camera and SuperGlue model

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors