Recurrent Neural Networks to Correct Satellite Image Classification Maps
This addresses the need for precise object outlining in aerial and satellite imagery, which is crucial for applications like urban planning or environmental monitoring, representing an incremental improvement over existing enhancement algorithms.
The paper tackles the problem of poor object localization in satellite image classification maps produced by CNNs by proposing a method to directly learn an iterative enhancement process using recurrent neural networks (RNNs), resulting in significantly improved map quality.
While initially devised for image categorization, convolutional neural networks (CNNs) are being increasingly used for the pixelwise semantic labeling of images. However, the proper nature of the most common CNN architectures makes them good at recognizing but poor at localizing objects precisely. This problem is magnified in the context of aerial and satellite image labeling, where a spatially fine object outlining is of paramount importance. Different iterative enhancement algorithms have been presented in the literature to progressively improve the coarse CNN outputs, seeking to sharpen object boundaries around real image edges. However, one must carefully design, choose and tune such algorithms. Instead, our goal is to directly learn the iterative process itself. For this, we formulate a generic iterative enhancement process inspired from partial differential equations, and observe that it can be expressed as a recurrent neural network (RNN). Consequently, we train such a network from manually labeled data for our enhancement task. In a series of experiments we show that our RNN effectively learns an iterative process that significantly improves the quality of satellite image classification maps.