Instructions

Using a model file for binary classification of single image data, the data format needs to match the specified model file. The returned result type is RasterRDD. Binary classification of images, which uses deep learning algorithms to determine whether the pixels in the image belong to an interest category, is usually used to obtain a single land class with obvious characteristics such as roads, rivers, and buildings. The result type is RasterRDD, and the pixel values are 1 and 0, respectively. A value of 1 indicates that the pixel belongs to an interest category. The following figure shows the binary classification analysis results of the original image overlay:

Parameter Specification