I would like to know what is the process that allows Object Detection. Example: I train the machine to recognize the cars and the machine recognizes and geo-references all the cars in the photo.
How can I do it?
Thank you for your interest in the Orfeo Toolbox.
To answer your question, I will need more information. Indeed, Object Detection on Remote Sensing images is a vast field, and there are many ways to accomplish what you are asking for. Do you think of a specific approach?
Since you are talking about “training the machine”, I guess you would be interested in this page of the CookBook. It presents all the OTBapplications used to do Machine Learning. For instance, TrainImagesClassifier is useful to train a classifier from multiple pairs of images and training vector data. You could also look at ImageClassifier, it can perform an image classification based on a model file produced by the TrainImagesClassifier. The combination of those two OTBapplication can meet your needs, if you set the parameters correctly.
I hope it helps.