# US 12,762,855 B2 · Systems and methods for calculating water resources using aerial imaging of crop lands > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,762,855 B2 - Document ID: US12762855B2 - Status: Granted - Kind: B2 - Application number: 17/840,497 - Filed: 2022-06-14 - Published: 2022-12-22 - Granted: 2026-10-13 - Earliest claimed: 2021-06-16 - Applicants: Ipq Pty Ltd - Inventors: Thomas Henry Rooney - Web page: https://patentgrep.com/patent/US12762855B2 ## CPC - G06V 20/188: {Vegetation} - G06V 10/267: {by performing operations on regions, e.g. growing, shrinking or watersheds} - G06V 10/764: using classification, e.g. of video objects - G06V 10/82: using neural networks - G06V 20/13: Satellite images - G06V 20/17: taken from planes or by drones - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING ## Abstract Systems and methods for determining groundwater levels based upon crop classification are provided. A set of aerial images are collected via satellite, manned aircraft or drones. They are filtered by a time domain, and a sufficiently high-resolution image is selected. If there isn't an image with sufficient resolution, a series of lower resolution images may be combined to generate a ‘fused’ image suitable for analysis. The image is then subjected to pre-processing. Crop boundaries within the image are determined, and areas outside of the crop boundary are masked off. The resulting image is subjected to a sliding window algorithm to generate discrete “patches” of the image suitable for analysis by a trained neural network. The neural network generated a classification for the crop. This data may be combined with surface water data, precipitation data, and weather pattern data to determine groundwater levels. ## Claims ### Claim 1 A method for classifying crops using aerial imagery comprising: collecting a plurality of aerial images for a given region; selecting at least one image of the plurality of images for analysis; determining a crop boundary within the at least one image; masking areas outside the crop boundary to generate a contiguous crop area; applying a sliding window algorithm to the contiguous crop area to generate a plurality of patch images; classifying the patch images using a neural network; and determining a crop type for the contiguous crop area by aggregating the patch image classifications; and calculating groundwater in an aquifer associated with the given region using the determined crop type, wherein the calculating accounts for evaporation within the given region based upon weather patterns, and crop water requirements under the weather patterns. ### Claim 2 (depends on 1) The method of claim 1, wherein the classifying the patch images includes a classification and a confidence level. ### Claim 3 (depends on 2) The method of claim 2, wherein the determining the crop type includes selecting the classifications for the patch images with the highest confidence levels. ### Claim 4 (depends on 2) The method of claim 2, wherein the determining the crop type includes selecting the classifications for the patch images that are most frequent. ### Claim 5 (depends on 1) The method of claim 1, wherein the selecting the at least one image includes: filtering the plurality of images by a time domain; determining if a sufficient resolution image is available; when the sufficient resolution image is available, selecting said sufficient resolution image; and when no sufficient resolution image is available, fusing lower resolution images into an amalgamated image. ### Claim 6 (depends on 1) The method of claim 1, wherein the aerial images are at least one of visible light images, ultraviolet light images and LiDAR images. ### Claim 7 (depends on 1) The method of claim 1, wherein the aerial images are collected by at least one of a satellite, a drone, or an aircraft. ### Claim 8 (depends on 1) The method of claim 1, wherein the calculating the groundwater in the aquafer includes combining the deter-mined crop type with surface water flows and precipitation. ### Claim 9 A computer program product embodied in a non-transitory storage medium, which when executed on a computer system performs the steps of: collecting a plurality of aerial images for a given region; selecting at least one image of the plurality of images for analysis; determining a crop boundary within the at least one image; masking areas outside the crop boundary to generate a contiguous crop area; applying a sliding window algorithm to the contiguous crop area to generate a plurality of patch images; classifying the patch images using a neural network; and determining a crop type for the contiguous crop area by aggregating the patch image classifications; and calculating groundwater in an aquifer associated with the given region using the determined crop type, wherein the calculating accounts for evaporation within the given region based upon weather patterns, and crop water requirements under the weather patterns. ### Claim 10 (depends on 9) The computer program of claim 9, wherein the classifying the patch images includes a classification and a confidence level. ### Claim 11 (depends on 10) The computer program of claim 10, wherein the determining the crop type includes selecting the classifications for the patch images with the highest confidence levels. ### Claim 12 (depends on 10) The computer program of claim 10, wherein the determining the crop type includes selecting the classifications for the patch images that are most frequent. ### Claim 13 (depends on 9) The computer program product of claim 9, wherein the selecting the at least one image includes: filtering the plurality of images by a time domain; determining if a sufficient resolution image is available; when the sufficient resolution image is available, selecting said sufficient resolution image; and when no sufficient resolution image is available, fusing lower resolution images into an amalgamated image. ### Claim 14 (depends on 9) The computer program product of claim 9, wherein the aerial images are at least one of visible light images, ultraviolet light images and LiDAR images. ### Claim 15 (depends on 9) The computer program product of claim 9, wherein the aerial images are collected by at least one of a satellite, a drone, or an aircraft. ### Claim 16 (depends on 9) The computer program product of claim 9, wherein the calculating the groundwater in the aquafer includes combining the determined crop type with surface water flows and precipitation.