Method for creating generative time series dataset for change detection in remote sensing

A method of training a remote sensing change detection model, the method comprising:

• downloading, by a user input device, aerial and Earth satellite images of high resolution or greater, wherein the high resolution or greater is a pixel size of 0.3 m/pixel or less;
• selecting, by the user input device, a static image from among the aerial and satellite images;
• generating, by multi-core processors that cyclically perform fused multiply-add operations for a depth map generator, an original depth map of the static image, where the depth map is an image that contains information relating to distance of surfaces of objects from a viewpoint;
• creating, by processing circuitry configured with a semantic map creator, a semantic map of the static image, wherein the semantic map includes annotated objects in the static image;
• generating, by the processing circuitry configured with a change simulator that iteratively determines a candidate area for change simulation, a change depth map and a change mask, wherein the change mask focuses on one or more objects removed from the static image;
• generating, by multi-core processors that cyclically perform fused multiply-add operations for a tile image generator, a pair of change tiles and a mask using the change depth map, wherein the pair of change tiles includes a post-change tile and a pre-change tile;
• iteratively validating, by validation processing circuitry, the pair of change tiles and the mask to obtain a validated pair of change tiles and masks;
• wherein the validating includes comparing each object in the pair of change tiles to each of a plurality of ground truth objects,
• wherein a change tile is invalid when a object in the change tile is a distorted object or results in an impossible context such that the object does not substantially match any of the plurality of ground truth objects,
• wherein when the pair of change tiles are found invalid, store an identifier to indicate that the pair of change tiles is invalid in a rejected list;
• and
• training, by the multi-core processors, the remote sensing change detection model with a database of a plurality of validated pairs of change tiles and masks.

A non-transitory computer-readable storage medium including computer executable instructions, wherein the instructions, when executed by an AI workstation, cause the computer to perform a method for generating land use area change images, the method comprising:

• downloading, by a user input device, aerial and Earth satellite images of high resolution or greater, wherein the high resolution or greater is a pixel size of 0.3 m/pixel or less;
• selecting, by the user input device, a static image from among the aerial and satellite images;
• generating, by multi-core processors that cyclically perform fused multiply-add operations for a depth map generator, an original depth map of the static image, where the depth map is an image that contains information relating to distance of surfaces of objects from a viewpoint;
• creating, by processing circuitry configured with a semantic map creator, a semantic map of the static image, wherein the semantic map includes annotated objects in the static image;
• generating, by the processing circuitry configured with a change simulator that iteratively determines a candidate area for change simulation, a change depth map and a change mask, wherein the change mask focuses on one or more objects removed from the static image;
• generating, by multi-core processors that cyclically perform fused multiply-add operations for a tile image generator, a pair of change tiles and a mask using the change depth map, wherein the pair of change tiles includes a post-change tile and a pre-change tile;
• iteratively validating, by validation processing circuitry, the pair of change tiles and the mask to obtain a validated pair of change tiles and a validated change mask,
• wherein the validating includes comparing each object in the pair of change tiles to a plurality of ground truth objects,
• wherein a change tile is invalid when a object for the change tile is a distorted object or results in an impossible context such that the object does not substantially match any of the plurality of ground truth objects,
• wherein when the pair of change tiles are found invalid, store an identifier to indicate that the pair of change tiles is invalid in a rejected list;
• and
• storing the validated pair of change tiles and the validated mask in a database.