The test targets for the perception function of autonomous vehicles are crucial equipment for evaluating the reliability and safety of automotive ADAS systems. However, due to the lack of effective regulatory technical means, a large number of obviously deformed and damaged targets are still used repeatedly and for extended periods until they are completely damaged during testing. The test results obtained using poor-quality or even unqualified targets pose potential safety risks that cannot be ignored for subsequent open road testing and consumer use, highlighting the urgent need to establish practical and effective quantitative calibration methods for the optical characteristics of these targets. In response to this urgent need, this paper studies and establishes a calibration method for the optical spatial distribution reflection characteristics of typical humanoid targets. Furthermore, based on a novel three-dimensional polyhedral standard body, industrial cameras, and a bidirectional reflection distribution function analysis algorithm using the photographic method, a set of traceable on-site calibration devices is constructed. This has positive significance for establishing and improving relevant regulatory technical means, evaluation standards, and technical specifications, as well as exploring traceable technology for supervising the entire life cycle of test target quality.
This paper reported the research and establishment of a set of spectral retro-reflection standard measurement and calibration device. Based on the physical definition and measurement principle of spectral retro-reflection standard, a calibration device that can provide stable and reliable measurement results in the spectral band of 380nm - 780nm and under key retro-reflection measurement geometric conditions (incident angle: -4° - 30°, observation angle: 0.2° - 1°) was constructed and implemented through the precise optical path design integrating a stable light source and optical fiber spectrometer. Experiments showed that the device performed well in measurement accuracy and consistency. Measurement verification experiments were carried out with the national retro-reflection coefficient calibration device, and the optimal consistency measurement results for the same sample under characteristic conditions were better than 1.0%. This work has positive significance for further promoting the application of retroreflective materials in transportation, safety, intelligent connected vehicles and other fields.
In computer vision tasks, various types of objects exhibit distinct characteristics in images. By learning and synthesizing the commonalities present in the training set, the neural network effectively performs tasks associated with diverse objects. However, when the training set is incomplete—specifically, when certain classes are missing—it becomes challenging for the network to learn the features of these absent classes during testing. Consequently, the coverage of the training data must be evaluated prior to training the network. This study uses synthetic aperture radar (SAR) aircraft detection as an example to illustrate the importance of evaluating dataset coverage, introduce evaluation methods, and propose solutions for incomplete dataset. Variations in SAR target features occur when the relative observation angle of SAR changes, causing changes in the brightness of the target scattering points. SAR images can exhibit significantly different characteristics even for similar targets. Based on this characteristic, the aircraft in SAR images are classified into eight angle-based classes. If the training set includes fewer than eight angle types for aircraft (at least one), the network will be unable to detect aircraft from all eight angles in the test set. To tackle potential issues arising from incomplete training sets, the following solutions are proposed: Firstly, a clustering algorithm is employed to classify the labeled data more accurately by considering the differences in the heat maps of various feature data. Next, the average heat map is extracted for each data class, overlaid, and compared with the average heat map of the complete test set to identify any missing data types. Finally, the training set is supplemented with the appropriate data based on the identified missing data types. Experimental results using partial data from the SAR-AIRcraft-1.0 dataset demonstrate the effectiveness of the proposed method.
Accurate traceability of LiDAR(Light Detection and Ranging) is still a difficult issue in the autonomous driving industry, as it serves as crucial 'eyes' for perception. To address this issue, a device was developed for absolute measurement of reflectance at small angles near θ/θ reflection condition of LiDAR. Through the fine design of the small angle optical path, the problem of mutual occlusion between the light source and the detector due to the actual size under this condition was optimized, the accurate value of the incident and reflected radiant flux were obtained, and the absolute measurements of the reflectance with the minimum angle better than 0.1°/-0.1° were realized. To verify the accuracy and reliability of the device, one 1 -inch diameter UV-enhanced aluminum film plane mirror and one 1-inch diameter silver film plane mirror were used. Absolute reflectance measurements were conducted at light source wavelengths of 532nm and 905nm, with reflection angles of 12°/-12°and 45°/-45°. The measured results were then compared with their calibrated values, achieving an optimal relative deviation of 0.4%, thereby providing preliminary validation of the device's measurement accuracy and reliability. Meanwhile, the light source and detection part of this device are planned to be extended to 1550 nm, providing better support for LiDAR reflection traceability in the autonomous driving industry.
As a pair of "wise eyes" for autonomous vehicles to perceive the external environment, Lidar (Light detecting and Ranging) plays a crucial role in detecting target characteristics in driving scenarios. To ensure the accuracy and reliability of Lidar, precise measurement of target reflectance under θ/θ reflection conditions is an indispensable step. Determining the reflectance value of targets under θ/θ reflection conditions is a critical part of completing Lidar calibration and traceability. When designing the θ/θ optical path, a significant challenge lies in achieving almost perfect overlap between the lighting and detecting paths while ensuring system compactness and measurement accuracy. Fiber optic spectrometers, known for their fast and accurate measurement capabilities, can be directly applied to measure target reflectance. Therefore, combining a compact θ/θ reflection optical path with a fiber optic spectrometer is key to achieving small-angle reflectance measurements for Lidar, marking an important step towards improving the calibration and traceability chain. For Lidar under θ/θ reflection conditions, various "N+1" lighting/detecting combined optical paths based on fiber bundles have been designed. Simulation analysis of these designs has been conducted using ray-tracing methods, comparing the uniformity and optical flux efficiency of the tested samples. The results indicate that when N=6, the uniformity and incident flux efficiency are optimal. Based on the simulation results, a "6+1" lighting/detecting fiber optic spectrometer has been developed, and actual measurements have been performed on a standard diffuser. The measurement data shows that the angular accuracy under 0/0 and 45/45 conditions is better than 0.1°, and the optimal relative error of the reflection measurement results in the 905nm laser wavelength was less than 0.5%. This meets the requirements for on-site measurements and is significant for further improving the Lidar measurement traceability chain.
The 2-D galvo scanners refer to motorized mirror mounts and systems specifically designed for applications involving laser-beam steering or scanning. They excel in swiftly manipulating small laser beams, offering exceptional levels of accuracy and precision. Functioning as dynamic electro-optical components, galvo scanners employ a rotatable mirror with low inertia to accurately position a laser beam with a high degree of precision and repeatability. In a series of applications, achieving a repeatability of less than 2 μrad is critical for galvo scanner. However, the galvo scanner system could suffer from optical and control errors because of lacking researchers’ intervention. These errors may give rise to deviations between the actual motion of the galvo scanner and the desired motion, then impede the precise positioning and control of the beam. Hence, an exceedingly precise and accurate calibration of the galvo scanner is imperative to attain high-precision control. To realize the standard accuracy of 0.3 mm in applications, current methods still keep imperfections. To evaluate the accuracy of the galvanometer's scanning beam location across a specified voltage range and the variations in its scanning capabilities with signal frequency, this research introduces a high-precision calibration measurement method for the 2-D galvo scanner. This method serves as a valuable resource for future studies on galvanometer scanner calibration and trajectory matching algorithms, establishing a research foundation for the eventual implementation of uniform light field illumination by galvanometers.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.