In remote sensing applications, the accuracy of the results is very important in order to present accurate and useful results to the user. Therefore, if the product results are not evaluated, their ability is limited and the product cannot be used with certainty. Accuracy assessment for classification requires the comparison of the land use and land cover classification at every pixel in the image with the reference or ground truth information.
Classification accuracy is then assessed using an error matrix (confusion matrix) and comparing the proportions of times that the classes are correctly predicted to the number of times they are misclassified (Congalton, 1991: Foody, 2002). To measure the classification accuracy, there is need to generate an error matrix table. Error matrix is a square array of numbers laid out in rows and columns that expresses the number of sample units assigned to a particular category relative to the actual category as verified in the field. The columns normally represent the reference data, while the rows indicate the classification generated from remotely sensed data.
Basically, in error matrix table consist of a few accuracy assessment some of which are producer’s accuracy (omission error), user’s accuracy (commission error), overall accuracy and Kappa coefficient. Assessing accuracy of classification using class proportions and kappa statistics multinomial distribution and assumes a large sample normality, respectively which are typically easier to satisfy (Agresti, 2002).
Congalton and Plourde (2002) indicated that ground truth data collection, classification scheme, sampling scheme, spatial auto correlation, sample size and sample unit are the factors that must be considered in order to properly generate an error matrix. In error matrix, the correct sample point in a class will be calculated based on the number of points of that class in the ground truth data.
In addition, the assessment also based on Kappa statistic. Kappa statistics are used to test the relationship between the observed and the classification. Kappa statistic value is between +1 (perfect agreement), 0 (no agreement) and -1 (complete disagreement). Equation 1 shows the kappa coefficient Lillesand et al., (2000).
Knowledge of current land use, agriculture, and also information on their changing proportion, which is required by law, planners, and local government officials is important to determined preferably by land use and land cover evaluation. The results can support policy makers to achieve long-term sustainability of land and water resources and its impact on climate change.
Land use planning involves making decisions for the use of land and natural resources. The goal is to make use the land that is productive and the implementation is sustainable to meet the current and the future need. Land use planning is done using different scales, that is based on the planning stages, namely the national, state, district, and local field survey (Yusrizal, 2008).
Land use is an important element in the management of town and country planning. There are many types of land use information needed for planning and monitoring as an area of land use, they include availability of housing, employment, economic and transport links. Without this information on land use, JPBD cannot provide planning services as defined in the Town and Country Planning Act, 1976 (Act 172), 2001.
Following this amendment is very important to manage land use information efficiently and comprehensively. Table 2.5 show the land use classification that is used by the local plan. All the land use map that are produce must follow each color and code number for maintain the standardization among all agencies.