Laserfiche WebLink
MEMO 89.2 <br />Kiefer 2000). Only crop types classified by the MLC algorithm were evaluated in the accuracy <br />assessment. Manually assigned classes, such as `Vegetables', were not included in this assessment. <br />Vegetables represent less than 2 percent of the irrigated acreage in Division 1. <br />Reference Parcels <br />Irrigated Crop <br />Classification <br />Overall, Producer', <br />and User' <br />Accuracies, Kappa <br />Statistics <br />Error Matrix <br />Figure 17. Irrigated Crops Accuracy Assessment Procedure <br />As with irrigated lands, the accuracy assessment procedure was used iteratively to refine the irrigated <br />crops classification. Figure 18 shows overall accuracy and Kappa Index of Agreement (KIA) of four <br />iterations of the irrigated crops classification refinement process. The higher accuracy of the `Initial Crop <br />Classification' is noticeable when compared to the `Initial Irrigated Lands Classification' in Figure 16. <br />This can be explained in part by the fact that most sources of confusion were eliminated in the irrigated <br />lands masking process. For example, riparian grass could have been confused with `Alfalfa'. The overall <br />accuracy achieved in the `Final Crop Classification' was 95 percent. This represents an improvement of <br />about 6 percent for overall accuracy, and 10 percent for overall KIA over the `Initial Crop Type. Error <br />matrices and corresponding derived accuracy assessment parameters for all four irrigated crop <br />classifications shown in Figure 18 are presented in Appendix D. <br />Page 26 of 45 ~R~versfde FecAnotogy, fnc. <br />4'JaYer Resources Errgi~ecr:np an~i CansuFlrnp <br />