Systems and methods for analyzing remote sensing imagery
Disclosed systems and methods relate to remote sensing, deep learning, and object detection. Some embodiments relate to machine learning for object detection, which includes, for example, identifying a class of pixel in a target image and generating a label image based on a parameter set. Other embodiments relate to machine learning for geometry extraction, which includes, for example, determining heights of one or more regions in a target image and determining a geometric object property in a target image. Yet other embodiments relate to machine learning for alignment, which includes, for example, aligning images via direct or indirect estimation of transformation parameters.
1. A method of determining a geometric object property in a target image, comprising:
receiving the target image, a geometric object corresponding to a portion of the target image, and a set of parameters comprising an orientation of an illumination source associated with the target image; and
determining, using a regression model, the geometric object property from the target image based on the geometric object and the set of parameters.
2. The method of claim 1 , wherein the target image is one of red-green-blue, grayscale, panchromatic, infrared, ultraviolet, multi-spectral, or hyperspectral image.
3. The method of claim 1 , wherein the geometric object is representative of a roof.
4. The method of claim 3 , wherein the geometric object property comprises a roof material.
5. The method of claim 3 , wherein the geometric object property comprises a roof area.
6. The method of claim 3 , wherein the geometric object property comprises a roof pitch.
7. The method of claim 3 , further comprising determining a roof condition based on the geometric object property.
8. The method of claim 1 , wherein the set of parameters further comprises at least one of: a position of a target sensor, a target time, a target date, a target latitude, or a target longitude.
9. The method of claim 1 , wherein the geometric object is represented using a binary mask.
10. The method of claim 9 , wherein the binary mask is generated by semantically segmenting the target image.
11. The method of claim 1 , wherein the geometric object is aligned with the target image.
12. The method of claim 1 , wherein the target image comprises satellite imagery.
13. A system, comprising non-volatile computer readable media configured to determine a geometric object property of a geometric object in a target image by:
Receiving the target image, the geometric object corresponding to a portion of the target image, and an orientation of an illumination source associated with the target image; and
determining the geometric object property, using a classifier, based on the target image, the geometric object, and the orientation of the illumination source.
14. The system of claim 13 , wherein the target image comprises a geospatial image.
15. The system of claim 13 , wherein the geometric object is representative of a roof and the geometric object property is a dominant pitch of the roof.
16. The system of claim 13 , further configured to detect vegetation within the target image.
17. The system of claim 13 , wherein the target image is not a stereo image.
18. The system of claim 13 , wherein the classifier is trained using a training image comprising a structure and an associated heatmap depicting the structure.
19. The system of claim 13 , wherein the classifier is trained using a training image and an associated heatmap comprising labels for multiple geometric objects.
20. The system of claim 19 , wherein the multiple geometric objects comprise a roof and vegetation.