Construction stage detection using satellite or aerial imagery
Methods, non-transitory computer-readable storage media, and computer or computer systems directed to detecting, analyzing, and tracking stages of housing construction using satellite or aerial imagery in combination with a machine learned model are described.
1. A method comprising:
selecting a geographic area of interest on a map or satellite or aerial image;
causing a satellite or aerial image or portion thereof corresponding to the geographic area of interest to be sent as input for a machine learned model trained with a set of satellite or aerial images having features characteristic of housing at four different stages of construction and corresponding labels representing such stages; and
receiving one or more output from the machine learned model, the output comprising one or more predictions of the stages of construction determined for the features in the satellite or aerial image or portion thereof;
wherein the four different stages of construction are slab, foundation, under construction, and completed;
wherein the selecting and causing are based upon input from a Geographic Information System application;
wherein the output comprises text formatted and standardized for input into the Geographic Information System application; and
wherein the text formatted and standardized for input into the Geographic Information System application comprises:
a feature name;
a label for the feature representing a prediction chosen from slab, foundation, under construction, and completed;
a decimal probability that the label for the feature is correct; and
a centroid and polygonal boundary of the feature, both expressed as geographical coordinates.
2. The method of claim 1 , wherein the input comprises one or more location information chosen from information comprising city, county, state, zip code, geographic coordinates and tax parcel number.
3. The method of claim 1 , wherein the input comprises providing an outline surrounding the geographic area of interest.
4. The method of claim 1 , wherein the machine learned model is a trained Convolutional Neural Network (CNN).
5. The method of claim 1 , wherein the selecting, causing, and receiving are performed by one or more processors.