Back to results
US 11,536,843 B2Granted
De-jitter of point cloud data for target recognition
AssigneeThe Boeing Company
InventorDennis J. Yelton
Life of this patent
Estimating the term…
Abstract
Jitter is removed from point cloud data of a target by fitting the data to 3-D models of possible targets. The point cloud data is de-jittered as a group by shifting the point cloud data in its coordinate system until a minimum fit error is observed between the shifted data and a 3-D model under analysis. Different 3-D models may be evaluated in succession until a 3-D model is identified that has the least fit error. The 3-D model with the least fit error most likely represents the identity of the target.
Claims21
A method comprising:
- (a) generating lidar point cloud data of a target;
- (b) selecting a 3-D model from a plurality of 3-D models of possible targets stored in memory;
- (c) mapping the selected 3-D model into a spatial coordinate system of the lidar point cloud data;
- (d) calculating a fit error between the lidar point cloud data and the selected 3-D model;
- (e) repetitively shifting the lidar point cloud data as a group in the spatial coordinate system and re-calculating the fit error until a minimum fit error for the selected 3-D model is identified;
- (f) repeating steps b-e for each of the 3-D models;
- (g) identifying the 3-D model having the least fit error;
- and
- (h) identifying the target based on the 3-D model having the least fit error.
6 more elements
A method, comprising:
- storing a plurality of 3-D models of possible targets in memory;
- generating point cloud data of an unidentified target by ranging the unidentified target with a flash ranging device;
- for each 3-D model of the plurality of 3-D models, performing the steps of:
- mapping the 3-D model into a spatial coordinate system of the point cloud data;
- calculating a fit error between the point cloud data and the 3-D model;
- and
- repetitively shifting the point cloud data as a group in the spatial coordinate system and re-calculating the fit error until a minimum fit error for the 3-D model is identified;
- identifying the 3-D model having the least fit error;
- and
- identifying the unidentified target based on the 3-D model having the least fit error.
7 more elements
A flash lidar system, comprising:
- a laser configured to generate laser light;
- a transmitter configured to illuminate a target with one or more pulses of the laser light;
- a receiver configured to detect the laser light reflected from the target and to generate lidar point cloud data of the target;
- a memory configured to store a plurality of 3-D models of possible targets;
- and
- a processor configured, for each 3-D model of the 3-D models, to map the 3-D model into a spatial coordinate system of the lidar point cloud data, to calculate a fit error between the lidar point cloud data and the 3-D model, and to repetitively shift the lidar point cloud data as a group in the spatial coordinate system and re-calculate the fit error until a minimum fit error for the 3-D model is identified,
- wherein the processor is further configured to identify the 3-D model having the least fit error, and to identify the target based on the 3-D model having the least fit error.
4 more elements
Dependencies are inferred from the claim text.