Single-view calibration of a catadioptric camera based on a theodolite model.

The catadioptric panoramic imaging system may provide 360° panoramic imaging by employing the convex surface of a quadric surface with rotational symmetry as the reflector, which effectively compensates for the disadvantages of the narrow field of view in typical camera systems. First, this paper proposes a theodolite-based catadioptric camera image model based on the rotational symmetry of a catadioptric camera mirror, which simplifies the 2D modeling problem to a 1D problem. Simultaneously, the equivalence of the theodolite imaging model and the standard spherical imaging model also is demonstrated in this work. Second, this paper presents a method to calibrate the theodolite model parameters using only a single view and explains the calculation of model parameter initialization and iterative optimization steps in detail. Then, this paper demonstrates how to calibrate the theodolite model parameters using only a single view, as well as how to calculate the model parameter initialization and iterative optimization steps. Finally, simulation experiments and actual experiments have confirmed the correctness and stability of the method. The experimental results show that the average and standard deviation of the back-projection error are, respectively, 0.1983125 pixels and 0.0006153 pixels, in this model. We believe the theodolite model suggested in this paper offers a viable approach to catadioptric camera image modeling.

Calibration of a Catadioptric System and 3D Reconstruction Based on Surface Structured Light.

In response to the problem of the small field of vision in 3D reconstruction, a 3D reconstruction system based on a catadioptric camera and projector was built by introducing a traditional camera to calibrate the catadioptric camera and projector system. Firstly, the intrinsic parameters of the camera and the traditional camera are calibrated separately. Then, the calibration of the projection system is accomplished by the traditional camera. Secondly, the coordinate system is introduced to calculate, respectively, the position of the catadioptric camera and projector in the coordinate system, and the position relationship between the coordinate systems of the catadioptric camera and the projector is obtained. Finally, the projector is used to project the structured light fringe to realize the reconstruction using a catadioptric camera. The experimental results show that the reconstruction error is 0.75 mm and the relative error is 0.0068 for a target of about 1 m. The calibration method and reconstruction method proposed in this paper can guarantee the ideal geometric reconstruction accuracy.

High-Resolution Drone Detection Based on Background Difference and SAG-YOLOv5s.

To solve the problem of low accuracy and slow speed of drone detection in high-resolution images with fixed cameras, we propose a detection method combining background difference and lightweight network SAG-YOLOv5s. First, background difference is used to extract potential drone targets in high-resolution images, eliminating most of the background to reduce computational overhead. Secondly, the Ghost module and SimAM attention mechanism are introduced on the basis of YOLOv5s to reduce the total number of model parameters and improve feature extraction, and α-DIoU loss is used to replace the original DIoU loss to improve the accuracy of bounding box regression. Finally, to verify the effectiveness of our method, a high-resolution drone dataset is made based on the public data set. Experimental results show that the detection accuracy of the proposed method reaches 97.6%, 24.3 percentage points higher than that of YOLOv5s, and the detection speed in 4K video reaches 13.2 FPS, which meets the actual demand and is significantly better than similar algorithms. It achieves a good balance between detection accuracy and detection speed and provides a method benchmark for high-resolution drone detection under a fixed camera.

Pose Estimation of Omnidirectional Camera with Improved EPnP Algorithm.

The omnidirectional camera, having the advantage of broadening the field of view, realizes 360° imaging in the horizontal direction. Due to light reflection from the mirror surface, the collinearity relation is altered and the imaged scene has severe nonlinear distortions. This makes it more difficult to estimate the pose of the omnidirectional camera. To solve this problem, we derive the mapping from omnidirectional camera to traditional camera and propose an omnidirectional camera linear imaging model. Based on the linear imaging model, we improve the EPnP algorithm to calculate the omnidirectional camera pose. To validate the proposed solution, we conducted simulations and physical experiments. Results show that the algorithm has a good performance in resisting noise.

Improved difference model applied in the Fourier-transform-based integration method based on Taylor theory.

The two-dimensional Fourier-transform-based integration algorithm is widely used in shape or wavefront reconstruction from gradients. However, its reconstruction accuracy is limited by the truncation error of the difference model. The truncation error is affected by the distribution of the sampling points. It increases when the sampling points are unevenly distributed and arranged irregularly. For improving, a novel way to calculate the difference is proposed based on Taylor expansion theory of binary functions. The first-order partial derivative terms are used to estimate the second- and third-order partial derivative terms for reducing the truncation error. The proposed difference model is applied to Fourier-transform-based integration. The reconstruction results show that it can get better results when the sampling points are irregularly distributed.

Fast and non-iterative zonal estimation for the non-rectangular data in the transparent surface reconstruction from polarization analysis.

In the method of surface reconstruction from polarization, the reconstructed area is generally non-rectangular and contains a large number of sampling points. There is a difficulty that the coefficient matrix in front of the height vector changes with the shape of the measured data when using the zonal estimation. The traditional iterative approaches consume more time for the reconstruction of this type of data. This paper presents a non-iterative zonal estimation to reduce the computing time and to accurately reconstruct the surface. The index vector is created according to the positions of both the valid and invalid elements in the difference and gradient matrices. It is used to obtain the coefficient matrix corresponding to the general data. The heights in the non-rectangular area are calculated non-iteratively by the least squares method. At the same time, the sparse matrix is applied for handling the large-scale data quickly. The simulation and the experiment are designed to verify the feasibility of the proposed method. The results show that the proposed method is highly efficient and accurate in the reconstruction of the non-rectangular data.