Detection and Mapping of Kochia Plants and Patches using High-Resolution Ground Imagery and Satellite Data: Application of Machine Learning

The presence of Kochia weed can be harmful to crop production. It grow well in harsh conditions and resist popular herbicides such as glyphosate. causes chemical stress crops spreads quickly, forming large patches. Detecting early is crucial for effective control. However, it hard identify its stages closely resembles early-stage crops. Moreover, seeds spread from neighbouring farms through water or wind, making detection at both field regional levels necessary. Currently, object-based methods are used the level, but there still a lack literature on mapping level. Our research proposes methodology accurately detecting, localizing quantifying plants fields using high-resolution RGB imagery. We also explore potential detecting patches level satellite approach uses semantic segmentation techniques process geotagged images, allowing us quantify individual field. To ensure accurate detection, we have established minimum density threshold based images. This enables distinguish spectral signature imagery pixels with high Kochia. label geo-locations where exceeds value. method has 99% accuracy rate multi-spectral 40%. model trained in-field mean intersection over union value up 0.8606. These results suggest pixel-level performed more if pixel than 40% mix. study highlights data farm management. help prevent damage successful