Begin Single-image depth estimation holds significant application value in the field of computer vision. Traditional methods have certain limitations in accurately locating distant regions in images and predicting the precision of object edges. Therefore, this paper proposes a depth estimation method from a single image using fusion of multi-scale features. Firstly, we construct a feature extraction module that combines Transformer and multi-scale feature fusion mechanisms. This module effectively captures features at different scales and levels in images while focusing on global features, resulting in a more comprehensive and precise feature representation. Secondly, we design a Scene Feature Encoding Module that utilizes dilated convolution blocks and adaptive average pooling to extract context correlations between pixels in images, enabling effective recognition of distant regions. Finally, we design a Strip Spatial Perception Module to refine the perception of depth changes, thereby enhancing overall estimation accuracy. Experimental results showcase the outstanding performance of this approach across diverse scenarios, proving its practicality and wide-ranging application prospects in the field of depth estimation from a single image.
Turbulence profile is an important parameter for characterization of atmospheric turbulence intensity at different altitudes. Based on generalized Hufnagel-Valley atmospheric turbulence model and measurement data of DCIM lidar, we first analyzed the difficulties encountered by analytical methods, then a numerical method based on particle swarm optimization algorithm was proposed to retrieval the unknown parameters of generalized Hufnagel-Valley model. To enhance the accuracy of high altitudes, whole layer isoplanatic angle was also as a constraint in the inversion. Moreover,we compared the particle swarm optimization algorithm and Levenberg-Marquardt algorithm in term of inversion accuracy.
In this paper, the Faster R-CNN algorithm and YOLOv3 algorithm are researched and practiced based on the remote sensing image data sets. Using the same data sets and hardware environment, it mainly evaluates the average accuracy and the time-consuming for detection of the target objects in the data sets. These algorithm evaluation indicators evaluate the relative applicability of the two algorithms in practical applications. The reasons are also analyzed for the deficiencies of the two algorithms in the target detection process. It is concluded that the Faster R-CNN algorithm is more suitable for practical applications that require higher target detection accuracy, and the YOLOv3 is more suitable for practical applications that require less time-consuming.
This paper uses the measured atmospheric coherence length profile data of DCIM lidar to analyze the effect of different regularization parameter selection strategies on the inversion of atmospheric turbulence profile. The criterions of L-curve, generalized cross-validation(GCV), quasi-optimal are used respectively, The inversion results is evaluated by signal-tonoise ratio(SNR) and root mean square error(RMSE). The results show that the GCV criterion perform more stable for various measurements than L-curve and quasi-optimal criterion.
We develop differential column image motion (DCIM) lidar for monitoring atmosphere refractive structure constant Cn2 profile. It is important to use an appropriate regularization method for DCIM lidar since the ill-posedness of the integral equation between the Cn2 profile and the measured r0 profile. In this paper, three typical regularization methods are studied to retrieve the Cn2 profiles from r0 profiles.The experiments illustrate that the Tikhonov method and truncated SVD method perform good performance, while damped SVD shows poorer inversion accuracy.
Differential column image motion lidar (DCIM lidar) can obtain the Fried’s transverse coherence length (r0) of different altitudes with a high spatial and temporal resolution. According to the integral equation of atmospheric coherence length (r0) of spherical wave, the refractive structure constant(C2n) profile can be retrieved from r0 profile. Aiming at improving the retrieval accuracy of atmospheric turbulence profile, noise reduction on r0 profile is implemented before inversion. Two methods of wavelet threshold and complementary ensemble empirical mode decomposition (CEEMD) are used to denoise r0 profile. The effects of denoised methods on r0 profile and C2n profile are investigated. The numeric simulations and experiments are both carried out to validate the two denoised methods. The results show that both the two methods can improve the signal-to-noise ratio (SNR) of atmospheric coherent length profile and reduce the recovered error of the atmospheric turbulence profile, and wavelet threshold method is superior to CEEMD method under different noise conditions.
Differential column image motion lidar (DCIM lidar) is a recent turbulence monitor for acquiring atmospheric turbulence profile based on active beacon. By imaging the differential column onto a CCD, DCIM lidar can obtain the Fried’s transverse coherence length (r0) of different altitudes with a high spatial and temporal resolution. Atmospheric turbulence profile can be recovered from r0 profile based on the integral relationship between r0 of spherical wave and the refractive structure constant (C2n ). In order to ensure the retrieved precision of atmospheric turbulence profile, singular value decomposition (SVD) is used to denoise r0 profile before inversion. The theory of DCIM lidar and SVD denoising is described. The Hankel matrix is constructed from the noisy signal and then the SVD is used to obtain the singular values. The rank reduction parameter is determined from the sharp variation of singular value curve. The denoised signal can be reconstructed by choosing the bigger singular values according to the rank reduction parameter. The numeric simulations and experiments are both carried out to validate the denoised method of SVD. The results show that the SVD can increase signal-to-noise ratio of r0 profile, thus enhancing the accuracy of the recovered atmospheric turbulence profile.
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.