Classifying objects with additional measurements

Patent metadata

Field	Value
Publication number	US-10915765-B2
Application number	US-201916456426-A
Country	US
Kind code	B2
Filing date	Jun 28, 2019
Priority date	Jan 3, 2017
Publication date	Feb 9, 2021
Grant date	Feb 9, 2021

Abstract

A vehicle-assistance system for classifying objects in a vehicle's surroundings is provided. The system may include at least one memory configured to store classification information for classifying a plurality of objects and at least one processor configured to receive, on a pixel-by-pixel basis, a plurality of measurements associated with LIDAR detection results. The measurements may include at least one of: a presence indication, a surface angle, object surface physical composition, and a reflectivity level. The at least one processor may also be configured to receive, on the pixel-by-pixel basis, at least one confidence level associated with each received measurement, and access the classification information. The at least one processor may further be configured to, based on the classification information and the received measurements with the at least one associated confidence level plurality, identify a of pixels as being associated with a particular object.

First claim

What is claimed is:

1. A vehicle-assistance system for identifying objects in a vehicle's surroundings, the system comprising:
at least one processor configured to:
receive point-cloud information originating from a LIDAR configured to project light toward the vehicle's surroundings, wherein the point-cloud information is associated with a plurality of data points and each data point includes indications of a three-dimensional location and angular information with respect to a reference plane;
construct, from the received point cloud information, a point cloud map of the vehicle's surroundings, wherein the point cloud map is indicative of a shape of a particular object in the vehicle's surroundings and of angular orientations of at least two surfaces of the particular object, wherein the angular orientations of at least two surfaces of the particular object include a slope of at least one of the at least two surfaces of the particular object;
access object-related classification information; and
identify the particular object based on the angular orientations of the at least two surfaces of the particular object from the point cloud map and the object-related classification information.

2. The vehicle-assistance system of claim 1 , wherein identifying the particular object includes classifying a plurality of pixels as being associated with the particular object.

3. The vehicle-assistance system of claim 1 , wherein identifying the particular object includes determining a type of the particular object.

4. The vehicle-assistance system of claim 1 , wherein the point cloud map is further indicative of a reflectivity level of different portions of the particular object.

5. The vehicle-assistance system of claim 1 , wherein each data point in the point cloud map includes measurements of at least two of: a presence indication, a surface angle, a reflectivity level, velocity, and ambient light.

6. The vehicle-assistance system of claim 5 , wherein the received point cloud information further includes confidence level for each of the measurements.

7. The vehicle-assistance system of claim 1 , wherein the at least one processor is further configured to determine a match by identifying a most likely three-dimensional representation in the classification information that corresponds to the information in the point cloud map.

8. The vehicle-assistance system of claim 1 , wherein when a certainty level is under a threshold, the at least one processor is further configured to request additional information from the LIDAR.

9. The vehicle-assistance system of claim 1 , wherein the at least one processor is further configured to determine that the identified particular object is an object of interest and to forward the LIDAR an indication of a region of interest that includes the identified particular object.

10. The vehicle-assistance system of claim 1 , wherein identifying the particular object is further based on the shape of the particular object from the point cloud map.

11. A non-transitory computer-readable storage medium storing instructions that, when executed by at least one processor, cause the at least one processor to perform a method for identifying objects in a vehicle's surroundings, the method comprising:
receiving point-cloud information originating from a LIDAR configured to project light toward the vehicle's surroundings, wherein the point-cloud information is associated with a plurality of data points and each data point includes indications of a three-dimensional location and angular information with respect to a reference plane;
constructing, from the received point cloud information, a point cloud map of the vehicle's surroundings, wherein the point cloud map is indicative of a shape of a particular object in the vehicle's surroundings and of angular orientations of at least two surfaces of the particular object, wherein the angular orientations of at least two surfaces of the particular object include a slope of at least one of the at least two surfaces of the particular object;
accessing object-related classification information; and
identifying the particular object based on the angular orientations of the at least two surfaces of the particular object from the point cloud map and the object-related classification information.

Assignees

• Innoviz Tech Ltd

Inventors

• Day Amir
• Zohar Guy
• Vlaiko Julian
• Osiroff Nir
• Menadeva Ovadya

Classifications

• G06V20/625

  License plates · CPC title

• G06V10/60

  relating to illumination properties, e.g. using a reflectance or lighting model · CPC title

• G06V10/82

  using neural networks · CPC title

• G06F18/256

  of results relating to different input data, e.g. multimodal recognition · CPC title

• G06F18/2414

  Smoothing the distance, e.g. radial basis function networks [RBFN] · CPC title