3D Object Tracking via 2D Imaging and Height Map Projection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current autonomous vehicle systems face challenges in accurately tracking objects in three-dimensional space using two-dimensional imaging data, particularly due to limitations in camera systems and three-dimensional imaging technologies like Lidar, which can miss objects or provide incomplete information.

Innovation Solution

A system that combines two-dimensional imaging data from cameras with height map data to transform objects into three-dimensional space, using modules like data receiving, two-dimensional to three-dimensional transformation, object identification, and visual classification to refine object dimensions and location, and track objects effectively.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If two-dimensional imaging data from cameras is used to track objects, then the system complexity is reduced and cost is lowered, but the accuracy of three-dimensional object location and detection precision deteriorate

Engineering Contradiction:
Improvesystem complexityVSAvoidobject location accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent applies dimensionality change by projecting two-dimensional image data onto a three-dimensional height map to recover depth information. The 2D image coordinates are transformed into 3D world coordinates by intersecting rays from the camera pose with the height map surface, enabling accurate 3D object localization without requiring complex 3D imaging sensors.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Measurement precision

If three-dimensional imaging technologies like Lidar are used, then object detection precision and three-dimensional space tracking are improved, but the system complexity and cost increase

Engineering Contradiction:
Improveobject detection precisionVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent introduces a height map as an intermediary data structure that mediates between 2D camera images and 3D spatial understanding. The height map stores ground surface elevation information, allowing the system to transform 2D image points into 3D world coordinates by finding the intersection of camera rays with the height map surface, thus achieving 3D tracking without direct 3D sensors.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Device complexity

If traditional object tracking methods are used, then the device complexity is low, but the reliability of object tracking in incomplete three-dimensional imaging data deteriorates

Engineering Contradiction:
Improvedevice complexityVSAvoidobject tracking reliability
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The patent applies preliminary action by pre-computing and storing a height map of the environment before object tracking begins. This height map contains advance information about ground surface elevations and obstacles, enabling the system to reliably transform 2D image data into accurate 3D object locations even when real-time 3D imaging data is incomplete or unavailable.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS10318822B2Object tracking
Publication Date: 2019.06.11 GM GLOBAL TECHNOLOGY OPERATIONS LLC
  • US10318822B2 patent drawing
  • US10318822B2 patent drawing
  • US10318822B2 patent drawing

AI summary

Methods and systems are provided for tracking an object. The system includes a data receiving module receiving two dimensional imaging data including an object and height map data correlating ground height and location. A two dimensions to three dimensions transformation module determines a location of the object in three dimensional space based on the two dimensional imaging data and the height map data. A tracking module tracks the object using the location of the object.