Model for mapping settlements
A programmable media includes a graphical processing unit in communication with a memory element. The graphical processing unit is configured to detect one or more settlement regions from a high resolution remote sensed image based on the execution of programming code. The graphical processing unit identifies one or more settlements through the execution of the programming code that executes a multi-instance learning algorithm that models portions of the high resolution remote sensed image. The identification is based on spectral bands transmitted by a satellite and on selected designations of the image patches.
1. A method of analyzing high resolution remote sensing imagery, comprising:
an integrated circuit that includes a microcontroller on a single chip executes the functions comprising:
receiving satellite imagery data;
automatically dividing the high resolution imagery into image patches;
acquiring and automatically processing training data that renders bag of Gaussian models;
processing a query window comprising a portion of the imagery data with the bag of Gaussian models to classify objects recorded in the satellite imagery data.
2. The method of claim 1 where the act of dividing the high resolution imagery is based on the pixel resolution of satellite imagery.
3. The method of claim 1 where the act of dividing the high resolution imagery is based on the average size of objects recorded in the satellite imagery data.
4. The method of claim 1 where the act of dividing the high resolution imagery is based on the number of image bands transmitted by a satellite.
5. The method of claim 1 further comprising designating selected image patches into discriminate classes.
6. The method of claim 5 where the act of designating comprises color coding the selected image patches.
7. The method of claim 1 further comprising generating multi-dimensional feature vectors for each training data rendering a training pixel in a training patch.
8. The method of claim 1 further comprising generating a plurality of two-dimensional planes based on the number of spectral bands transmitted by a satellite.
9. The method of claim 1 where the classification is based on a probabilistic distance between the query window with the bag of Gaussian models.
10. The method of claim 9 where the classification is based on a ranking.