Methods and apparatus for automatically defining computer-aided design files using machine learning, image analytics, and/or computer vision

An apparatus, comprising:

• a memory;
• and
• a processor operatively coupled to the memory, the processor configured to receive aerial data having a plurality of points arranged in a pattern, the processor configured to provide an indication associated with each point from the plurality of points as an input to a machine learning model to classify each point from the plurality of points into a set of categories from a plurality of categories,
• the processor configured to select, for each point from the plurality of points, a category from the set of categories for that point based on a predefined category hierarchy,
• the processor configured to, for each point from the plurality of points, identify a set of points from the plurality of points adjacent to that point in the pattern and having a common category from the plurality of categories of that point to define a shape from a plurality of shapes,
• the processor configured to generate a computer-aided design file using each shape from the plurality of shapes.

A method, comprising:

• training, at a first time, a machine learning model using a first aerial dataset having a first plurality of shapes, each shape from the first plurality of shapes (1) is labeled with a category from a plurality of categories and (2) includes a polyline boundary fully encompassing that shape and distinct from the polyline boundary for the remaining shapes from the first plurality of shapes;
• receiving, at a second time after the first time, a second aerial dataset having a plurality of points arranged in a pattern;
• providing an indication associated with each point from the plurality of points as an input to the machine learning model to classify each point from the plurality of points into a category from the plurality of categories;
• for each point from the plurality of points, identifying a set of points from the plurality of points adjacent to that point in the pattern and having a common category from the plurality of categories of that point to define a shape from a second plurality of shapes;
• defining a polyline boundary of each shape from the second plurality of shapes by analyzing with respect to a criterion a position of each point associated with a border of that shape relative to a position of at least one other point associated with the border of that shape;
• and
• generating a computer-aided design file using the polyline boundary of each shape from the second plurality of shapes.

A non-transitory processor-readable medium storing code representing instructions to be executed by a processor, the code comprising code to cause the processor to:

• receive aerial data having a plurality of points arranged in a pattern;
• filtering the aerial data to define a set of points from the plurality of points;
• identify, for each point from the set of points, a group of points adjacent to that point in the pattern to define a cluster from a plurality of clusters;
• define a set of clusters from the plurality of clusters based on at least one of a number of points from the set of points in each cluster from the set of clusters or each point from the set of points in each cluster from the set of clusters meeting an elevation criterion;
• define a shape from a plurality of shapes formed by at least one cluster from the set of clusters based at least in part on the at least one cluster from the set of clusters meeting a regularity criterion;
• generate a polyline boundary of each shape from the plurality of shapes by analyzing with respect to a criterion a position of each point associated with a border of that shape relative to a position of at least one other point associated with the border of that shape;
• and
• generate a computer-aided design file using the polyline boundary of each cluster from the set of clusters.