Coastal erosion is a major problem that has been worsened by climate change, along with other natural occurrences like droughts and marine flooding. Countries situated along coastlines are facing significant challenges when it comes to preserving their land and protecting their people and assets. To mitigate the damage caused by the encroachment of the sea on land, effective monitoring tools and methods are required. Several remote sensing techniques and methods have been developed to address these issues, including machine learning and deep learning methods. In this study, Objectoriented Analysis (OBIA), Pixel-Oriented Analysis (PBIA), and Convolutional Neural Network (CNN) methods are used to automatically detect and extract the Greater Libreville coastline, based on Pléiades very high-resolution satellite image dating from 2022. Three test areas were chosen there and then extracted. The first zone is located in the north of the municipality of Akanda (marked by the presence of small coastal cliffs). The second zone is located in the commune of Libreville (sandy beach). The third zone is located in the municipality of Owendo (artificialized beach, mostly muddy). The images of the three zones were the subject of a classification based on the methods mentioned above. Results of proposed methodologies showed competitive Overall Accuracy (OA) values obtained with OBIA method and the CNN model. However, the OBIA method using Random Forest algorithm (RF) achieved the highest accuracy rates, which reached 95%, 90%, and 80% for the three test areas respectively.
Urban green land plays a special significant role in our life. Many previous studies have already proven the feasibility of satellite imagery processing for urban green land detection, and many classification techniques were tested for this purpose. In this paper, two methods of Machine learning, such as Artificial Neural Networks (ANN) and Random Forest (RF) were tested on a series of very high spatial resolution satellite imagery to classify, highlight urban green lands and eventually study their change in three years. Firstly, we present Multilayer Perceptron (MLP), a simple feed-forward neural network which consists of two steps, SLIC superpixel segmentation and image classification through MLP trained model. Secondly, RF technique was applied and compared to the previous classification. Then the classification results were interpreted at the final stage. These methods are performed on satellite images Pléiades of the city of Brest (3 images acquired in 2016, 2017 and 2018), these multispectral images are already preprocessed and all pan-sharpened to 50 cm spatial resolution.
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.