RADAR-CAMERA OBJECT DETECTION

Organization Name

ford global technologies, llc

Inventor(s)

Yunfei Long of East Lansing MI (US)

Daniel Morris of Okemos MI (US)

Abhinav Kumar of East Lansing MI (US)

Xiaoming Liu of Okemos MI (US)

Marcos Paul Gerardo Castro of Mountainview CA (US)

Punarjay Chakravarty of Campbell CA (US)

Praveen Narayanan of Santa Clara CA (US)

RADAR-CAMERA OBJECT DETECTION - A simplified explanation of the abstract

This abstract first appeared for US patent application 20240264276 titled 'RADAR-CAMERA OBJECT DETECTION

Simplified Explanation: This patent application describes a computer system that uses radar data and camera data to generate a combined set of object features and confidence scores.

The computer system includes a processor and memory.
The memory contains instructions for generating radar data by projecting radar returns onto an image plane based on camera and radar sensor parameters.
The system uses an image/radar convolutional neural network to process the data.
Features are transferred from the image channel to the radar channel at multiple stages.
Object features and confidence scores are determined separately for the image and radar channels.
The system combines image object features with radar object features using a weighted sum.

Key Features and Innovation:

Integration of radar and camera data for object detection.
Use of convolutional neural network for processing.
Separate determination of object features and confidence scores for radar and image data.
Combination of features from both data sources for improved accuracy.

Potential Applications:

Autonomous vehicles for enhanced object detection.
Surveillance systems for improved monitoring capabilities.
Robotics for better understanding of the environment.

Problems Solved:

Improved object detection accuracy.
Enhanced data fusion from multiple sources.
Better understanding of the scene for various applications.

Benefits:

Increased accuracy in object detection.
Enhanced situational awareness.
Improved decision-making capabilities for autonomous systems.

Commercial Applications:

Autonomous vehicle technology for safer transportation.
Surveillance systems for enhanced security measures.
Robotics for more efficient and effective operations.

Questions about the Technology 1. How does the weighted sum method improve the combination of image and radar object features? 2. What are the potential challenges in integrating radar and camera data for object detection?

Frequently Updated Research: Ongoing research in the field of sensor fusion and object detection algorithms may provide further advancements in this technology.

Original Abstract Submitted

a computer that includes a processor and a memory, the memory including instructions executable by the processor to generate radar data by projecting radar returns of objects within a scene onto an image plane of camera data of the scene based on extrinsic and intrinsic parameters of a camera and extrinsic parameters of a radar sensor to generate the radar data. the image data can be received at an image channel of an image/radar convolutional neural network (cnn) and receive the radar data at a radar channel of the image/radar cnn, wherein features are transferred from the image channel to the radar channel at multiple stages image object features and image confidence scores can be determined by the image channel, and radar object features and radar confidences by the radar channel. the image object features can be combined with the radar object features using a weighted sum.