This paper compares of pixel- and object-based techniques for mapping wild oat weed patches in wheatfields using multi-spectral QuickBird satellite imagery for site-specific weed management. The researchwas conducted at two levels: (1) at the field level, on 11 and 15 individual infested wheat fields in 2006 and2008, respectively, and (2) on a broader level, by analysing the entire 2006 and 2008 images. To evaluatethe wild oat patches mapping at the field level, both pixel- and object-based image analyses were testedwith six classification algorithms: Parallelepipeds (P), Mahalanobis Distance (MD), Maximum Likelihood(ML), Spectral Angle Mapper (SAM), Support Vector Machine (SVM) and Decision Tree (DT). The resultsshowed that weed patches could be accurately detected with both analyses obtaining global accuraciesbetween 80% and 99% for most of the fields. The MD and SVM classifiers were the most accurate forboth the pixel- and object-based images from 2006 and 2008, respectively. In the broad-scale analysis,all of the wheat fields were identified in the imagery using a multiresolution hierarchical segmentationbased on two scales. The first segmentation scale was classified using the MD and ML algorithms todiscriminate wheat fields from other land uses. Accuracies greater than 85% were obtained for MD and88% for ML for both imagery. A hierarchical analysis was then performed with the second segmentationscale, increasing the accuracies to 93% and 91% for 2006 and 2008 imagery, respectively. Finally, based onthe most accurate results obtained in the field-level study, pixel-based classifications using the MD, MLand SVM algorithms were applied to the wheat fields identified. The results of these broad-level analysesshowed that wild oat patches were accurately discriminated in all the wheat fields present in the entireimages with accuracies greater than 91% for all the classifiers tested.

This work was partially financed by the Spanish Ministry of Science and Innovation through projects AGL2008-04670-CO3-03 and AGL2011-30442-CO2-01 (FEDER). JAE-Doc CSIC FEDER Program financed the work of J.M. Pena.

Peer Reviewed