Anomaly detection
Disclosed herein are methods and systems for anomaly detection within a field using images or sensor data. An anomaly detection process determines whether a pixel or a cluster of pixels fit within calculated or modeled expectations to identify an anomaly. The process can further analyze the anomaly to determine a likely cause. Example anomaly detection techniques account for intra-field inhomogeneity and for non-crop regions. Example anomaly detection techniques also can identify anomalous regions.
1. A method for detecting anomalies in homogenous plant growth, the method comprising:
computing first pixel statistics of pixels of a sensed image of land;
generating a pixel mask using the first pixel statistics;
applying the pixel mask to the sensed image to generate a masked image;
computing second pixel statistics using the masked image;
identifying a pixel clump within the masked image using the second pixel statistics, the pixel clump representing an anomalous area within the land; and
generating an anomaly map using the pixel clump.
2. The method of claim 1 , further comprising predicting a source of anomaly of the pixel clump, and wherein the step of generating the anomaly map uses the predicted source of anomaly.
3. The method of claim 1 , further comprising:
determining a topographical position of the pixel clump; and
determining a likely source of anomaly of the pixel clump using the topographical position of the pixel clump.
4. The method of claim 1 , further comprising predicting a source of anomaly of the pixel clump by determining a spectral signature for the pixel clump.
5. The method of claim 4 , wherein the step of predicting the source of anomaly of the pixel clump further comprises determining a topographical position of the pixel clump.
6. The method of claim 5 , wherein the step of predicting the source of anomaly of the pixel clump further comprises comparing the spectral signature for the pixel clump against a topographical wetness index.
7. The method of claim 4 , wherein the step of predicting the source of anomaly of the pixel clump further comprises comparing the spectral signature for the pixel clump against a reflectance threshold.
8. The method of claim 4 , wherein the step of predicting the source of anomaly of the pixel clump further comprises determining the pixel clump has a greenness value that exceeds a threshold.
9. The method of claim 4 , wherein the step of predicting the source of anomaly of the pixel clump further comprises determining a ratio of red to near-infrared on the spectral index for the pixel clump.
10. The method of claim 1 , wherein the step of generating a pixel mask using the first pixel statistics comprises adding a pixel of the sensed image to the pixel mask responsive to determining that the pixel deviates from a mean pixel value by a predetermined threshold.
11. The method of claim 1 , wherein the step of identifying a pixel clump comprises adding a pixel of the masked image to the pixel clump responsive to determining that the pixel deviates from a mean pixel value by a predetermined threshold.
12. The method of claim 1 , wherein the first or second pixel statistics includes a mean pixel value, a median pixel value, standard deviation statistics, or pixel values relative to a threshold.
13. The method of claim 1 , wherein the step of generating an anomaly map uses field boundary data.
14. The method of claim 1 , further comprising using the second pixels statistics to identify areas of the land having non-anomalous plant variation.
15. A method of detecting anomalous areas within land, the method comprising:
stratifying a sensed image of land into strata, the strata representing zones that each have at least one substantially homogenous characteristic;
applying a filter to the sensed image that masks individual pixels of clusters of pixels of the sensed image, wherein a pixel is masked responsive to determining that the pixel deviates from a predetermined threshold pixel value;
calculating pixel statistics of the filtered sensed image according to the strata; and
using the pixel statistics to isolate one or more clumps of pixels of the filtered sensed image, each clump representing an anomalous area within the land.
16. The method of claim 15 , wherein the step of using pixel statistics to isolate one or more clumps of pixels comprises adding a pixel to a pixel clump when the pixel deviates from another predetermined threshold pixel value.
17. The method of claim 15 , further comprising classifying the anomaly of each pixel clump using a topographical position of a respective pixel clump.
18. The method of claim 17 , wherein classifying the anomaly of each pixel clump further comprises using a spectral signature of a respective pixel clump.
19. The method of claim 15 , further comprising generating an anomaly map with the isolated one or more pixel clumps.
20. The method of claim 19 , wherein the step of generating an anomaly map uses field boundary data.