Source: 3D Vision Workshop
Add a small assistant: dddvision, note: direction + school/company + nickname, pull you into the group. At the end of the article, industry subdivisions are attached
This article describes a new FPP system (Laser Projection Profiling System) calibration method that utilizes an auxiliary camera to skip the projector calibration steps required in traditional methods. By using a secondary camera during calibration, and then removing it, the authors achieved a single camera and single projector configuration that remained simple when reconstructing geometry. The advantages of this approach over traditional methods are that it saves the time required to obtain calibration data because only a one-way fringe pattern needs to be projected, it is not limited to a specific type of projector, it can be adapted to a wide range of laser projection systems, and it is possible to calibrate systems that use rough illuminators. Experimental results show that this new method has the same accuracy as the traditional method when reconstructing geometric shapes.
Let's read about this work together~
论文题目：Low-cost adaptive obstacle avoidance trajectory control for express delivery drone
作者:Yanhui Zhang, Caisheng Wei等
作者机构：School of Aeronautics and Astronautics, Zhejiang University等
Paper link: https://arxiv.org/pdf/2403.19956.pdf
Structured light projection technology is a representative active method for 3D reconstruction, but many researchers face challenges in the complex projector calibration process. To address this complexity, we used an additional camera, tentatively called a secondary camera, to eliminate the need for projector calibration. The secondary camera helps to construct reasonable model equations that enable the generation of world coordinates based on absolute phase information. Once calibration is complete, the secondary camera can be removed, mitigating occlusion issues and allowing the system to maintain its compact single-camera, single-projector design. Our approach not only solves the common problem of calibrating projectors in digital fringe projection systems, but also enhances the feasibility of 3D imaging systems that utilize fringe projection in various shapes without the need for a complex projector calibration process.
The result of the three-dimensional reconstruction of the statue. (a) A textured image of the statue. (b) Laser-projected image of the statue. (c) Three-dimensional reconstruction results of the statue taken using the proposed method calibration system.
The result of the reconstruction of the statue, photographed using a plastic crevice lighting system. (a) Textured image of the statue. (b) Striped projection image of the statue. These streaks are generated by plastic gaps and a light source (in this case, a headlight). (c) Three-dimensional reconstruction of the statue taken using a plastic crevice lighting system. The system was calibrated using the proposed method.
- The proposed method introduces a novel method for calibration of structured light systems, which significantly improves efficiency without compromising accuracy.
- The proposed method incorporates an auxiliary camera to facilitate calibration, effectively adapting to both digital and non-digital illumination sources.
- The proposed method reduces the time and resources required for structured light system calibration, making advanced applications easier to implement and more cost-effective.
- The proposed method was validated by extensive experimental analysis.
The basic principle of this paper is to propose a new system calibration method, called the quasi-calibration method, for the complex projector calibration problem that requires complex projector calibration in traditional structured light systems. Traditional structured light systems use a projector to generate a fringe pattern and capture the reflected signals with a camera for 3D reconstruction. However, calibrating a projector is a complex and expensive process that limits the application and cost-effectiveness of the system.
The key idea of this approach is to construct 3D geometry using traditional stereo vision methods without the need for projector calibration by introducing a secondary camera into the system. During the calibration process, the main camera and the auxiliary camera are first calibrated with standard stereo vision, and then the auxiliary camera is used to obtain the 3D geometry information of the calibration plate. Then, the relationship between the phase value and the world coordinates is established by simulating a plane model and a pixel-level calibration model. Finally, the system is calibrated and applied by measuring with the main camera and digital projector to reconstruct the 3D geometry of the target object based on the established relationships.
The advantage of this method is that it does not require a complex calibration process for the projector, greatly simplifies the design and use of the system, and can be adapted to the application under different equipment conditions. Through experimental verification, this method can achieve almost the same 3D reconstruction accuracy as traditional techniques, while simplifying the calibration process, making more advanced applications more feasible and cost-effective.
The main purpose of this experiment is to verify the effectiveness of the proposed calibration method for FPP system. Two CCD cameras and a digital light processing (DLP) projector were used in the experiment. The experimental procedure is as follows:
The primary and secondary cameras were calibrated using standard stereo vision methods. By taking images of the calibration plates at 60 different positions, the internal and external parameters of the two cameras were obtained.
Stereo cameras and projectors were used to reconstruct the 3D geometry of the calibration plate. The projector projects a pattern of vertical stripes on the calibration plate, which is captured by two cameras, generates a phase map, and calculates the parallax. The projection pattern includes 18 phase offset patterns and 7 grayscale patterns.
The geometric fitting of the calibration plate and the determination of the ideal plane model are carried out. Using the fitted planar model, the relationship between the (xc, yc, zc) values and the Φ values of each pixel was established, and a rational model was adopted.
Use the main camera and projector to estimate the geometry. Repeat the pattern projection process in step 2 to acquire an image of the calibration plate and estimate the geometry based on the established pixel relationships.
Compared to traditional FPP calibration methods. Images of the calibration plates at the same position were acquired, the system was calibrated using the traditional method, and compared with the proposed method. The traditional method requires a pattern in both horizontal and vertical directions, while the proposed method only requires a stripe pattern in one direction.
Experimental results show that the proposed method has the same reconstruction accuracy as the traditional method, and can be applied to different types of projectors, such as digital light processing projectors and LED plastic gap lighting systems.
This paper describes a novel calibration method for FPP systems that skips the projector calibration step by utilizing an auxiliary camera. Later, the secondary camera was removed, allowing the system to maintain a simple single-camera and single-projector configuration for reconstructing geometry. Traditional FPP calibration methods require projecting horizontal and vertical fringe patterns. The new method requires only a one-way fringe pattern, which will halve the time it takes to collect calibration data per pose. In the experiments, the FPP system was calibrated using a vertical-only pattern, while ensuring the same level of accuracy in reconstructing the geometry. In addition, unlike traditional FPP calibration methods, which are not limited to DLP projectors, any laser projection system can be calibrated, even if it does not follow a pinhole model. In the second experiment, it was demonstrated that the new method can calibrate FPP systems using rough illuminators consisting of LEDs and plastic gaps.
This article is only for academic sharing, if there is any infringement, please contact to delete the article.
3D Vision Workshop Exchange Group
At present, we have established multiple communities in the direction of 3D vision, including 2D computer vision, large models, industrial 3D vision, SLAM, autonomous driving, 3D reconstruction, drones, etc., and the subdivisions include:
2D Computer Vision: Image Classification/Segmentation, Target/Detection, Medical Imaging, GAN, OCR, 2D Defect Detection, Remote Sensing Mapping, Super-Resolution, Face Detection, Behavior Recognition, Model Quantification Pruning, Transfer Learning, Human Pose Estimation, etc
Large models: NLP, CV, ASR, generative adversarial models, reinforcement learning models, dialogue models, etc
Industrial 3D vision: camera calibration, stereo matching, 3D point cloud, structured light, robotic arm grasping, defect detection, 6D pose estimation, phase deflection, Halcon, photogrammetry, array camera, photometric stereo vision, etc.
SLAM: visual SLAM, laser SLAM, semantic SLAM, filtering algorithm, multi-sensor fusion, multi-sensor calibration, dynamic SLAM, MOT SLAM, NeRF SLAM, robot navigation, etc.
Autonomous driving: depth estimation, Transformer, millimeter-wave, lidar, visual camera sensors, multi-sensor calibration, multi-sensor fusion, autonomous driving integrated group, etc., 3D object detection, path planning, trajectory prediction, 3D point cloud segmentation, model deployment, lane line detection, Occupancy, target tracking, etc.
3D reconstruction: 3DGS, NeRF, multi-view geometry, OpenMVS, MVSNet, colmap, texture mapping, etc
Unmanned aerial vehicles: quadrotor modeling, unmanned aerial vehicle flight control, etc
In addition to these, there are also exchange groups such as job search, hardware selection, visual product landing, the latest papers, the latest 3D vision products, and 3D vision industry news
Add a small assistant: dddvision, note: research direction + school/company + nickname (such as 3D point cloud + Tsinghua + Little Strawberry), pull you into the group.
3D Vision Workshop Knowledge Planet
3DGS, NeRF, Structured Light, Phase Deflection, Robotic Arm Grabbing, Point Cloud Practice, Open3D, Defect Detection, BEV Perception, Occupancy, Transformer, Model Deployment, 3D Object Detection, Depth Estimation, Multi-Sensor Calibration, Planning and Control, UAV Simulation, 3D Vision C++, 3D Vision python, dToF, Camera Calibration, ROS2, Robot Control Planning, LeGo-LAOM, Multi-modal fusion SLAM, LOAM-SLAM, indoor and outdoor SLAM, VINS-Fusion, ORB-SLAM3, MVSNet 3D