In many scene classification applications, the variety of surface objects, high within-category diversity and between-category similarity carry challenges for the classification Framework. Most of CNN-based classification methods only extract image features from a single network layer, which may cause the completed image information difficult to extract in complex scenes. We propose a novel transfer deep combined convolutional activations (TDCCA) to integrate both the low-level and high-level features. Extensive comparative experiments are conducted on UC Merced database, Aerial Image database and NWPU-RESISC45 database. The results reveal that our proposed TDCCA achieves higher experimental accuracies than other up-to-date popular methods.
Image is the important source of information for modern war. And making effective use of the information from the image would take full advantage of the reconnaissance capability. When images captured under fog, they are vulnerable to suspended particles in the atmosphere of the light scattering, absorption and other effects, and images suffer from quality degradation problems which lead to many difficulties for battlefield reconnaissance and recognition. Combining the dark and bright channel priors (bi-channel priors), the super-pixels are used as local regions, thus local transmission and atmospheric light values are estimated more reliably and efficiently. Furthermore, adaptive bi-channel priors are developed to rectify any incorrect estimation on transmission and atmospheric light values for both white and black pixels those fail to satisfy the assumptions of the bi-channel priors. Experimental results demonstrate that the white and black pixels on the restored UAV image are with excellent fidelity and the proposed method performs better for restoring images in terms of both quantitation and quality, and leads to great improvements in real-time defogging.
In order to realize detection and precise positioning of small-caliber visual optical system targets, according to the "cat eye effect" of photoelectric system, the research based on laser active reconnaissance precision detection technology was carried out. The effects of parameters such as laser emission power, receiving aperture and detection distance on the detection performance are simulated. The experiment has verified the ability to detect small-caliber targets under certain laser power. The experimental results show that the 10mm aperture visual optical system at 700m can be accurately detected under the condition of laser peak power of 1000W and laser divergence angle of 1.5mrad.
Unmanned aerial vehicles have been widely used in military and civil areas, which requires vision processing in explicit usage scenario. Existence of haze or fog can influence the context awareness capability of the aerial vehicles and makes affectation on target tasks. The captured images in hazy scenes suffer from degradation problems including poor contrast, color distortion, incomplete information, which lead to many difficulties in the follow-up processing. A simple and effective single image dehazing algorithm based on atmospheric scattering model and the optimum of image quality evaluation is proposed in this paper. Three image quality evaluation parameters: image entropy, standard deviation, and Fourier amplitude are combined to establish and the image quality evaluation function. On the basis of quality evaluation function, the image with the optimum of quality evaluation among the potential defogging images is chosen as the best result. Results show that this method has lower computational complexity, simplified operations and improved real-time performance.
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.