Localisation of an image in a known environment

I am currently trying to localize an image taken from a robot, I have the x axis value, y axis and angle the photographs was taken. What is the best way to go about doing this? It is a known environment so it isnt quite SLAM but similar.

Answer 1

[Shadaab Siddiqie]

From my understanding you want to find the location from an image. Here is a similar question answered previously.

[Benjamin Currie]

Hi Shadaab,

i dont think i worded my question very well. So basically, i have over 1 thousand images taken from a robot. I have performed semantic segmnetation on these images. Each image has an x, y and theta location from where the photo was taken. I want to use these x, y and theta values along with the information gathered from semantic segmentation ( number of blobs , how big the blobs are etc) and use this information to determine the location of the robot at a certain image based on this. I was thibking mauybe use an ANN?

[Shadaab Siddiqie]

Hey following is a link to monocular slam based algorithm that uses feature matching technique to localize the images, I think feature based approach for localization is better than blob based approach. Also i think you can extract features from segmentation. Then you can further correct x,y and theta based on correspondence between pose and localization using graph based slam.