Document images often contain critical textual information, and once manipulated or edited, they can cause significant harm to society when propagated on the internet. To date, although some works have achieved research results in the field of document image forgery detection, they have not fully explored edge features, which are important feature information in document images. In response to this research status, we propose a new document image forgery detection method based on multiscale edge similarity (MES). The proposed method employs multiscale difference of Gaussian (MDOG) to calculate the edge structure similarity, edge contrast similarity and edge information entropy at each pixel between the image to be detected and the reference image at different scales. These computed results are combined to generate an edge similarity distribution map, which is used to pinpoint the specific location of the forged area. Extensive experimental results demonstrate that the proposed method outperforms some of the current state-of-the-art forgery detection methods and shows robustness against several common postprocessing operations.
The automatic identification of grassland plants based on UAV remote sensing imagery is crucial for the management and conservation of grassland ecosystems. Grassland plants, characterized by small individual sizes, complex backgrounds, and dense distribution, pose a challenge for UAV image identification. An improved YOLOv8 algorithm is proposed in this study to enhance the detection effect of grassland plants from different angles and optimize the detection of small targets in complex backgrounds. Firstly, a rotation-invariant mechanism is introduced into the backbone to improve the model’s adaptation to grassland plant targets from various angles, thereby improving detection accuracy and robustness. Secondly, a self-attention mechanism is applied in the backbone to enhance the model’s understanding capability of correlations between different plants. Additionally, a multi-level feature extraction structure based on dilated convolution is employed in the neck to further enhance the model’s feature representation capability. Finally, the SlideLoss loss function is introduced to solve the problems of insufficient sample size and foreground-background detection impact. Experimental results demonstrate that the improved algorithm achieves remarkable results in the automatic identification of grassland plants in Inner Mongolia, validating the application prospects of UAV remote sensing technology in grassland plant species monitoring.
Convolutional neural networks (CNNs) are the mainstream model for extracting rich features in deep learning-driven studies on cloud detection for remote sensing images. However, due to the limitation of receptive fields in convolutional operations, a single CNN has limitations in mining global information and long-range dependencies, which affects the precision of cloud identification in intricate environments. To address the insufficient global feature and long-range dependency modeling of CNNs, a new CNN-Transformer network for cloud detection (CNN-TFCD) in remote sensing images is proposed in this paper. This method comprehensively utilizes strengths of both CNN and transformer to fully mine local spatial details, global contextual information, and long-range dependencies, to obtain comprehensive feature representations. The feature cascade mechanism is applied for multi-scale feature fusion, further enhancing the model's feature representation capability. Evaluations conducted on the Landsat-8 dataset demonstrate that the CNN-TFCD method achieves accurate and stable cloud detection.
Since 2008, straw burning has been vigorously banned in China. To investigate the impact of straw burning on AOD, this study used VIIRS nominal and high confidence fire point data and MODIS cropland land cover data to identify straw burning fire points. Additionally, by utilizing MOD04_3K aerosol products, the study analyzed the spatiotemporal correlation between straw burning during the summer and autumn harvest seasons and the aerosol optical deep (AOD) from 2012 to 2021. The results show that: (1) The distribution of straw burning fire points and ADO in Shandong Province exhibits significant seasonal characteristics, with AOD being heavily influenced by seasonal variations. (2) There is a linear relationship between the number of straw burning fire points and the seasonal mean AOD, with a correlation coefficient of 0.8145 in summer and 0.8907 in autumn. The correlation in autumn is slightly higher than that in summer. (3) The distribution of fire points coincides with the high AOD areas, mainly concentrated in the northern, western, southern, and central-eastern parts of Shandong Province, which are primarily wheat and corn planting regions. (4) The AOD in Shandong province showed a decreasing trend during the decade, and it was more obvious in the northern, western and central-eastern regions. These areas have the same spatial distribution as the regions where straw burning has decreased.
This paper presents a deep learning based method for inversion of Aerosol Optical Depth (AOD) from Landsat 8 OLI multi-band satellite data. Traditional aerosol inversion algorithms include the dark target method, the deep blue algorithm and other algorithms. The dark target method is good for areas with low surface reflectivity such as dense vegetation and water bodies, while the deep blue algorithm is usually used for inversion in areas with high reflectivity. Therefore, this study employs Deep belief networks (DBN) to learn the potential relationship between Landsat 8 OLI multi-band data and AOD. In this paper, Landsat 8 OLI remote sensing images with atmospheric AOD observations over the past 6 years (2013-2018) were collected as the training dataset. A deep confidence network structure is next designed for learning the mapping relationships from Landsat 8 OLI multi-band data to AOD. To evaluate the performance of the proposed method, we used a test set for validation. The experimental results show that the proposed deep learning-based method has higher accuracy and stability in AOD inversion compared to traditional methods.
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.