Trained machine learning model for estimating structure feature measurements

A system comprising:

• one or more processors;
• memory, wherein the memory stores one or more programs capable of execution by the one or more processors, and the one or more programs comprise instructions for:
• providing a plurality of training image sets, each training image set comprising one or more orthographic images depicting a real-world structure, wherein an orthographic image comprises a top-down view of the real-world structure;
• providing a plurality of training measurements, each training measurement corresponding to a real-world measurement of a feature of the real-world structure depicted in a respective training image set;
• scale normalizing each of the orthographic images, wherein scale normalizing an orthographic image comprises applying a pixel scalar to the orthographic image such that each pixel of the scale-normalized orthographic image corresponds to the same real-world metric distance of pixels across the orthographic images;
• and
• training a machine learning model using the scale-normalized orthographic images and the plurality of training measurements, wherein the trained machine learning model is configured to estimate a measurement of a feature of a real-world structure from an input image set.

One or more non-transitory computer-readable media storing instructions that when executed by one or more processors, cause the one or more processors to perform operations comprising:

• providing a plurality of training image sets, each training image set comprising one or more orthographic images depicting a real-world structure, wherein an orthographic image comprises a top-down view of the real-world structure;
• providing a plurality of training measurements, each training measurement corresponding to a real-world measurement of a feature of the real-world structure depicted in a respective training image set;
• scale normalizing each of the orthographic images, wherein scale normalizing an orthographic image comprises applying a pixel scalar to the orthographic image such that each pixel of the scale-normalized orthographic image corresponds to the same real-world metric distance of pixels across the orthographic images;
• and
• training a machine learning model using the scale-normalized orthographic images and the plurality of training measurements, wherein the trained machine learning model is configured to estimate a measurement of a feature of a real-world structure from an input image set.