Method and system for image capture and neural network training

A method for image capture and neural network training to classify a coastal area in a Remote Sensing (RS) image, comprising:

• capturing the RS image with an aerial sensor on an unmanned aerial vehicle;
• retrieving, from a memory, the RS image having a first resolution;
• processing, by a machine learning engine, the RS image through a multi-block residual attention-based Very Deep Super Resolution (VDSR) network to obtain a residual image, including dynamically adjusting weights in each block of the multi-block residual attention-based VDSR network having a series of convolutional layers with a residual connection and an attention mechanism, to enhance high-frequency details of the RS image;
• generating, by the machine learning engine, a high-resolution output image by adding the residual image to the RS image;
• generating, by the machine learning engine, a color feature map using the high-resolution output image;
• generating, by the machine learning engine, a resultant feature map by processing the high-resolution output image through a multimodal inception attention network (M2IAN) including a plurality of inception attention processes, each inception attention process of the plurality of inception attention processes including a spatial attention branch and a channel attention branch configured to identify special and channel features;
• combining, by the machine learning engine, the color feature map and the resultant feature map, including a depth-wise concatenation between the color feature map and the resultant feature map, to generate a composite feature image that captures subtle color variations in coastal areas;
• learning, by the machine learning engine, to classify the composite feature image into a plurality of coastal area categories;
• and
• displaying, by a display device, the composite feature image and classification category.