US 12,347,204 B2
Adaptive feature extraction to detect letters and edges on vehicle landing surfaces
Nikhil Gupta, Bengaluru (IN); and Shouvik Das, Bangalore (IN)
Assigned to Honeywell International Inc., Charlotte, NC (US)
Filed by Honeywell International Inc., Charlotte, NC (US)
Filed on Apr. 21, 2022, as Appl. No. 17/726,007.
Claims priority of application No. 202211011854 (IN), filed on Mar. 4, 2022.
Prior Publication US 2023/0282001 A1, Sep. 7, 2023
Int. Cl. G06V 20/56 (2022.01); B64D 45/08 (2006.01); B64D 47/08 (2006.01); G06T 7/13 (2017.01); G06V 10/40 (2022.01)
CPC G06V 20/56 (2022.01) [B64D 45/08 (2013.01); B64D 47/08 (2013.01); G06T 7/13 (2017.01); G06V 10/40 (2022.01)] 20 Claims
OG exemplary drawing
 
1. A vision-based landing system, comprising:
at least one processor onboard an aerial vehicle;
at least one vision sensor onboard the aerial vehicle and operatively coupled to the at least one processor;
one or more aiding sensors onboard the aerial vehicle and operatively coupled to the at least one processor; and
a data storage unit in operative communication with the at least one processor;
wherein the at least one processor hosts an adaptive feature extraction module, which has processor readable instructions to perform a method to detect edges on a landing area for the aerial vehicle, the method comprising:
capturing at least one image of the landing area with the at least one vision sensor, wherein the landing area includes a plurality of edge features;
selecting a first narrowed down region of interest in the at least one image to produce a first digital image;
calculating an estimated slope of an expected edge feature of the landing area;
calculating an expected gradient direction of the expected edge feature;
selecting a horizontal basis kernel for the expected edge feature along an expected horizontal gradient;
selecting a vertical basis kernel for the expected edge feature along an expected vertical gradient;
calculating a combined convolution kernel for the expected edge feature based on the horizontal and vertical basis kernels, the estimated slope of the expected edge feature, and the expected gradient direction of the expected edge feature;
performing a first convolution operation on the first digital image using the combined convolution kernel to produce a convolved image having first set of image pixels, which are binarized to obtain a resulting edge image; and
performing at least a second convolution operation on an entirety of the at least one image to produce a resulting edge feature image with extracted lines based on the plurality of edge features;
wherein a magnitude of values in the horizontal basis kernel and the vertical basis kernel are adapted to be changed according to an expected gradient strength and a dilution of the expected edge feature in pixels.