Automated geospatial image mosaic generation with automatic cutline generation
View Patent ↗Automatic generation of a mosaic comprising a plurality of geospatial images. An embodiment of the automatic mosaic generation may include automated source image selection that includes comparison of source images to base layer image to determine radiometric similar source images. Additionally, an embodiment of an automatic cutline generator may be provided to automatically determine a cutline when merging two images such that radiometric differences between the images along the cutline are reduced. In this regard, less perceivable outlines may be provided. Further still, an embodiment of a radiometric normalization module may be provided that may determine radiometric adjustments to source images to match certain properties of the base layer image. In some embodiments, when processing source images, the source images may be downsampled during a portion of the processing to reduce computational overhead. Additionally, some highly parallel computations may be performed by a GPU to further enhance performance.
1. A method for automatic cutline generation for merging at least two geospatial images to produce a composite image, the method comprising:
identifying at least a first geospatial image and a second geospatial image, wherein at least a portion of the first geospatial image and the second geospatial image overlap in an overlapping region;
obtaining a low resolution first geospatial image corresponding to the first geospatial image and a low resolution second geospatial image corresponding to the second geospatial image;
determining a low resolution cutline relative to adjacent pixels of the low resolution first geospatial image and the low resolution second geospatial image in the overlapping region, wherein the cutline is located between adjacent pixels from respective ones of the low resolution first geospatial image and the low resolution second geospatial image based on a radiometric difference therebetween;
expanding the low resolution cutline to define a cutline area in the overlapping region of the low resolution first and second images, wherein the cutline area is defined by cutline area boundaries;
applying the cutline area boundaries to the overlapping region of the first and second geospatial images to define a corresponding cutline area in the overlapping region of the first and second geospatial image; and
establishing a high resolution cutline relative to adjacent pixels of the first geospatial image and the second geospatial image in the cutline area, wherein the high resolution cutline is located between adjacent pixels from respective ones of the first geospatial image and the second geospatial image based on a radiometric difference therebetween in the cutline area.
2. The method of claim 1 , further comprising:
merging the first geospatial image and the second geospatial image to produce a composite image, wherein image data from the first geospatial image is provided on a first side of the cutline and image data from the second geospatial image is provided on an second side of the cutline opposite the first side.
3. The method of claim 2 , wherein the obtaining comprises downsampling the first geospatial image to produce the low resolution first geospatial image and downsampling the second geospatial image to produce the low resolution second geospatial image.
4. The method of claim 2 , wherein the downsampling comprises reducing the spatial resolution of the first image and the second image by at least a factor of two in both the vertical and horizontal directions.
5. The method of claim 2 , wherein the radiometric differences between adjacent pixels is determined utilizing a cost function that quantifies the radiometric difference between adjacent pixels from different corresponding images.
6. The method of claim 5 , wherein the cost function minimizes the radiometric differences between adjacent pixels from different images on opposite sides of the cutline.
7. The method of claim 6 , wherein the cost function comprises a max-flow min-cut cost function.
8. The method of claim 2 , wherein the expanding comprises encompassing a predetermined plurality of pixels on either side of the low resolution cutline to define the boundaries of the cutline area.
9. The method of claim 2 , wherein the cutline area comprises a subset of pixels of the first image and second image.
10. The method of claim 2 , wherein the first geospatial image partially covers an area of interest and the second geospatial image partially covers the area of interest, wherein the first geospatial image and the second geospatial image provide at least some unique coverage with respect to the area of interest.
11. The method of claim 1 , wherein more than two geospatial source images are to be merged, and wherein the method further comprises:
selecting the second geospatial source image to be merged with the first geospatial source image from a plurality of other geospatial source images, wherein the selecting comprises determining which of the plurality of other geospatial source images would contribute the most additional pixels to the composite image after the cutline has been established between a respective one of the plurality of other geospatial source images and the first geospatial source image.
12. The method of claim 11 , wherein the selecting step is repeated for each additional one of the plurality of other geospatial source images, wherein the first geospatial source image comprises a merged image comprising the first geospatial source image and each subsequent one of the plurality of other geospatial source images merged based on previous iterations of the selecting.
13. A system for generating a merged image comprising at least two geospatial images to produce a composite image, comprising:
an image database comprising at least a first geospatial image and a second geospatial image, wherein at least a portion of the first image and the second image overlap in an overlapping region;
a downsampling module operable to downsample each of the first geospatial image and the second geospatial image to generate a low resolution first geospatial image corresponding to the first geospatial image and a low resolution second geospatial image corresponding to the second geospatial image;
an automatic cutline generation module operable to:
determine a low resolution cutline relative to adjacent pixels of the low resolution first geospatial image and the low resolution second geospatial image in the overlapping region, wherein the cutline is located between adjacent pixels from respective ones of the low resolution first geospatial image and the low resolution second geospatial image based on radiometric differences therebetween in the overlapping region;
expand the low resolution cutline to a cutline area defined by cutline area boundaries;
apply the cutline area boundaries to the overlapping portion of the first geospatial image and the second geospatial image; and
establish a high resolution cutline relative to adjacent pixels of the first geospatial image and the second geospatial image in the cutline area, wherein the high resolution cutline is located between adjacent pixels from respective ones of the first geospatial image and the second geospatial image based on radiometric differences therebetween in the cutline area;
wherein the automatic cutline generation module is operable to output a merged image
wherein pixels on one side of the high resolution cutline comprise pixels from the first geospatial image and pixels on the other side of the high resolution cutline comprise pixels from the second geospatial image.
14. The system of claim 13 , wherein the automatic cutline generation module comprises a cost function for determining the radiometric difference between adjacent pixels, wherein the cost function quantifies the radiometric difference between adjacent pixels from different corresponding images.
15. The system of claim 14 , wherein the cost function is operable to determine the cutline so as to minimize the radiometric differences between adjacent pixels from different images on opposite sides of the cutline.
16. The system of claim 15 , wherein the cost function comprises a max-flow min-cut algorithm executable by a graphics processing unit of the automatic cutline generation module.
17. The system of claim 13 , wherein the expanding comprises encompassing a predetermined plurality of pixels on either side of the low resolution cutline to define the boundaries of the cutline area.
18. The system of claim 13 , wherein the cutline area comprises a subset of pixels of the first image and second image.
19. The system of claim 13 , wherein the first geospatial image partially covers an area of interest and the second geospatial image partially covers the area of interest, wherein the first geospatial image and the second geospatial image provide at least some unique coverage with respect to the area of interest.
20. The system of claim 13 , wherein more than two geospatial source images are to be merged, and wherein the automatic cutline generation module is further operable to select the second geospatial source image to be merged with the first geospatial source image from a plurality of other geospatial source images, wherein the selecting comprises determining which of the plurality of other geospatial source images would contribute the most additional pixels to the composite image after the cutline has been established between a respective one of the plurality of other geospatial source images and the first geospatial source image, and wherein the selecting step is repeated for each additional one of the plurality of other geospatial source images, wherein the first geospatial source image comprises a merged image comprising the first geospatial source image and each subsequent one of the plurality of other geospatial source images merged based on previous iterations of the selecting.