The invasion of Spartina alterniflora (S. alterniflora) has resulted in significant losses in the diversity of coastal ecosystems. However, the impact of seasonal changes on the accurate identification of S. alterniflora remains to be explored, which is of great significance due to its early monitoring and warning. In this study, S. alterniflora in Beihai, Guangxi, was selected as the research object. Unmanned aerial vehicles (UAVs) and deep convolutional neural networks (CNNs) were used to explore the identification of S. alterniflora in different seasons. Through comparative analysis, the ResNet50 model performed well in identifying S. alterniflora, with an F1-score of 96.40%. The phenological characteristics of S. alterniflora differ in different seasons. It is difficult to accurately monitor the annual S. alterniflora using only single-season data. For the monitoring of S. alterniflora throughout the year, the autumn-winter two-season model was selected from the perspective of time cost, the four-season model was selected from the perspective of identification performance, and the three-season model of summer, autumn and winter was selected from the perspective of time cost and identification performance. In addition, a method was developed to predict and evaluate the diffusion trend of S. alterniflora based on time series UAV images. Using the spring data to predict the diffusion trend of S. alterniflora in summer and autumn, the results showed that the highest recall reached 84.28%, the F1-score was higher than 70%, and most of the diffusion of S. alterniflora was correctly predicted. This study demonstrates the feasibility of distinguishing S. alterniflora from native plants in different seasons based on UAV and CNN recognition algorithms. The proposed diffusion early warning method reflects the actual diffusion of S. alterniflora to a certain extent, which is of great significance for the early management of invasive plants in coastal wetlands.
The traditional image quality assessment (IQA) methods are usually based on convolutional neural networks (CNNs). For these IQA methods using CNNs, limited by the feature size of the fully connected layer, the input image needs be tailored to a pre-defined size, which usually results in destroying the original structure and content of the input image and thus reduces the accuracy of the quality assessment. In this paper, a blind image quality assessment method (named CSPP-IQA), which is based on multi-scale spatial pyramid pooling, is proposed. CSPP-IQA allows inputting the original image when assessing the image quality without any image adjustment. Moreover, by facilitating the convolutional block attention module and image understanding module, CSPP-IQA achieved better accuracy, generalization and efficiency than traditional IQA methods. The result of experiments running on real-scene IQA datasets in this study verified the effectiveness and efficiency of CSPP-IQA.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.