News dynamics
Integration of RF, optoelectronic, embedded, automation and software technologies to provide dedicated equipment and solutions
On-line external reference calibration method without target based on common characteristics of millimeter wave radar and camera
Release time:
2021-11-09 10:29
Source:
Source
Sensor fusion is crucial for autonomous driving and autonomous robots, and meter-wave radar camera fusion systems are popular due to their complementary sensing capabilities. However, accurate calibration between these two sensors is more important to ensure effective fusion and improve overall system performance. The calibration between the two includes internal reference calibration and external reference calibration, the latter is particularly important for achieving accurate sensor fusion. Unfortunately, many target-based calibration methods require complex operating procedures and well-designed experimental conditions, which pose a challenge to researchers. In order to solve this problem, this paper introduces a new method that uses deep learning to extract common features from raw millimeter wave radar data (ie, range Doppler angle data) and camera images. Instead of representing these common features explicitly, our approach implicitly exploits these common features to match the same objects from two data sources, specifically, the extracted common features as-an example, demonstrating an online target-free calibration method between a meter-wave radar and a camera system. In order to improve the accuracy and robustness of calibration, RANSAC and L evenberg-Marquardt (LM) nonlinear optimization algorithms are used to derive the matrix, and the effectiveness and accuracy of the proposed method are proved by experiments in real environment.
Meter-wave radar and cameras are complementary sensing methods that are widely used in applications such as autonomous driving and robotics. Regardless of lighting and weather conditions, the meter wave radar can provide accurate distance, speed and angle information, while the camera can capture high-resolution visual information. The combination of meter-wave radar and camera improves perception and performs tasks such as object recognition, detection and tracking in a dynamic environment. The calibration between the meter-wave radar and the camera is essential for sensor fusion because of their different sensing principles, especially for determining their relative attitude. -- A promising method is to use deep learning to extract useful features from the original millimeter wave radar data, and explore the relationship between radar and image features to estimate the external transformation matrix
Recommended News
As communication bands rapidly shift toward millimeter waves (30 GHz–300 GHz), traditional Bias-Tees are facing three critical challenges. The BT110A, launched by Zhongwei Puyi, not only pushes the frequency upper limit to 110 GHz but also achieves domestic substitution with “military-grade reliability plus civilian‑grade costs,” making it a benchmark product that breaks the bottleneck in high‑frequency RF technology.
The Bias-T developed by Zhongwei Puyee can serve as a drop-in replacement for foreign SHF and MARKI products, with performance metrics superior to those of its competitors. Moreover, it has successfully completed all relevant environmental testing and validation.
5G rail transit, supporting the development of smart city rail.
At present, the operating mileage of urban rail transit in China ranks first in the world, and the number of cities and the length of lines that have built rail transit are in the forefront of the world. However, with the slowdown in the pace of urban rail transit construction, the pressure on the operation and maintenance management of existing lines is increasing. During the "14th Five-Year Plan" period, the focus of urban rail transit work has gradually shifted from construction to a stage of transformation and development with equal emphasis on operation and management.
"Neurospuye" high-speed data cable test solution
With the continuous development of 5G communication, cloud computing and the Internet of Things, the current transmission rate requirement for digital communication cables has reached 400Gbps and is developing to 800Gbps to 1.6Tbps, which also puts forward higher requirements for the shipment quality of high-speed transmission cables and high-speed cable assembly manufacturers, and requires better and more efficient cable detection schemes to meet.
On September 12, 2023 was the annual technology event of German science and technology (Keysight World Tech Day) held in Shanghai Pudong Kerry hotel. As a joint exhibition partner of the conference, we will discuss 5G/6G, auto driving and new energy, AI and data center technologies with global customers and partners.
Understanding 5G NTN Satellite Communication Technology in One Article
PCBA board testing is a key step to ensure that high-quality, high-stability and high-reliability PCBA products are delivered to customers, reduce defects in customers' hands, avoid after-sales, and strengthen customers' trust and the company's brand reputation. The following is a brief introduction to the testing method of PCBA board in PCBA processing industry
-
Address: Floor 10, Building 17, District 18, Fengtai District Headquarters Base, Beijing
Zip Code: 100070
Telephone:010-63718360
Email:sales@aumiwalker.com
Company website:www.aumiwalker.com
Enterprise WeChat
Taobao Store