Adaptive local scanning: A comprehensive and intelligent method for fast scanning of indiscrete objects

A pixel-by-pixel scanning that is usually performed by a single point-like sensor or probe is being widely used in the applications such as scanning probe microscopy techniques. Typically, their scanning time is several seconds to minutes long due to a raster scanning that needs to be conducted for capturing every single point on the surface of the sample area. To improve the scanning efficiency, recent research has been focused on investigating effective scanning patterns and methods. This work presents an adaptive local scanning method for efficiently sampling indiscrete objects like string-like one-piece connected objects under the microscopy. An initial scanning pattern is firstly investigated. Once the initial scanning reaches the object, an adaptive sinusoidal scanning method that can on-line adjust its scanning frequency and amplitude by predicting both the curvatures and the shape of the object is employed. The method also addresses scanning intersections and bifurcations associated with objects. Based on extensive implementations, it is validated that our method has high performance as it has high scanning efficiency and the scanned results match objects with high precision and high accuracy.