Edge optimization of semantic segmentation results is a challenging issue in remote sensing image processing. This paper proposes a semantic segmentation model guided by a block-in-block edge detection network named BIBED-Seg. This is a two-stage semantic segmentation model, where edges are extracted first and then segmented. We do two key works: The first work is edge detection, and we present BIBED, a block-in-block edge detection network, to extract the accurate boundary features. Here, the edge detection of multi-scale feature fusion is first realized by creating the block-in-block residual network structure and devising the multi-level loss function. Second, we add the channel and spatial attention module into the residual structure to improve high-resolution remote sensing images' boundary positioning and detection accuracy by focusing on their channel and spatial dimensions. Finally, we evaluate our method on ISPRS Potsdam and Vaihingen data sets and obtain ODS F-measure of 0.6671 and 0.7432, higher than other excellent edge detection methods. The second work is two-stage segmentation. First, the proposed BIBED is individually pre-trained, and subsequently, the pre-trained model is introduced into the entire segmentation network to extract boundary features. In the second segmentation stage, the edge detection network is used to constrain semantic segmentation results by loss cycles and feature bootstrapping. Our best model obtains the OA of 90.2%, 87.7%, and 81.5%, the IOU of 76.0%, 69.6%, and 61.3% on the ISPRS and WHDLD datasets, respectively.

<!-- *** Custom HTML *** --> Large intraclass variance and low interclass variance are among the most challenging problems in very high-resolution (VHR) image classification. Semantic segmentation constructed in a deep convolution neural network is used as a classification algorithm conducted via end-to-end training, which combines spectral–spatial features and context information. However, large-scale remote sensing images cannot be directly processed because they are limited by GPU memory and segmentation algorithm. At the same time, classification using single band combinations is also unsatisfactory due to the extraordinary complex features of VHR images. Therefore, a method is proposed based on multiple band combinations and patchwise scene analysis. A complex remote sensing image can be considered as into a combination of simple scenes from multiple patchwise images. And optimal band combinations of each patchwise image are selected according to their scene. The segmentation results of each patchwise image are merged to get the desired results according to geographical coordinates. Our method is validated on the ISPRS 2-D Semantic Labeling dataset of Potsdam, on which results competitive with the state-of-the-art are obtained. The proposed scheme has strong universality and can be used for large-scale high-resolution remote sensing image classification.

Extracting building instances from remote sensing images has various applications. However, existing instance segmentation methods have difficulty maintaining regular and angular building boundaries, corrupting the mask quality. This letter proposes a boundary-optimized instance segmentation method based on the direction field to cope with the low boundary quality. The proposed method develops a DF-Mask head to improve the mask quality of the primary instance segmentation network (Mask R-CNN) through boundary optimization. Within the DF-Mask head, first, a gated directional context-aware module (GDCAM) improves the mask features through the directional context and a designed gating mechanism. Then a direction field rectification module (DFRM) predicts the direction field and iteratively rectifies the mask. To the best of our knowledge, this is the first time that direction field rectification has been developed to the instance segmentation method and applied to building instance extraction. Experimental results on the Vaihingen and Massachusetts datasets suggest that our method outperforms the eight state-of-the-art methods. Compared with Mask R-CNN, our method improves APs by 4.7% and 2.2% on the two datasets, respectively.

Timely monitoring of marine aquaculture has considerable significance for marine ecological protection and maritime safety and security. Considering that supervised learning needs to rely on a large number of training samples and the characteristics of intensive and regular distribution of the laver aquaculture zone, in this paper, an inaccurate supervised classification model based on fully convolutional neural network and conditional random filed (FCN-CRF) is designed for the study of a laver aquaculture zone in Lianyungang, Jiangsu Province. The proposed model can extract the aquaculture zone and calculate the area and quantity of laver aquaculture net simultaneously. The FCN is used to extract the laver aquaculture zone by roughly making the training label. Then, the CRF is used to extract the isolated laver aquaculture net with high precision. The results show that the $kappa$ coefficient of the proposed model is 0.984, the $F_1$ is 0.99, and the recognition effect is outstanding. For label production, the fault tolerance rate is high and does not affect the final classification accuracy, thereby saving more label production time. The findings provide a data basis for future aquaculture yield estimation and offshore resource planning as well as technical support for marine ecological supervision and marine traffic management.

The classification of remote sensing images with high spatial resolution requires considerable training samples, but the process of sample making is slow and laborious. How to guarantee the accuracy of supervised classification under the condition of limited samples is an urgent problem to be solved in the field of supervised classification. For addressing this problem, we proposes an improved conditional generative adversarial network with edge feature (ECGAN) to augment limited training data for the classification of remote sensing images with high spatial resolution in this paper. On the basis of conditional generative adversarial network, feature factors of interclass boundaries and intraclass edges are added to networks, and an objective function with multiscale and multilevel features is constructed. The ISPRS Potsdam and Vaihingen remote sensing datasets are regarded as examples. Results indicate that the high-resolution remote sensing images generated by using the network proposed in this paper have abundant texture, accurate edges, and are highly similar to real images. The generated images are used to augment training samples, and an experiment for classifying high-resolution remote sensing images is conducted. The classification results of the proposed augmentation method perform better than that of the traditional sample augmentation method. We prove that ECGAN as a means of sample augmentation can effectively solves the problem that the classification effect is unideal when the supervised classification sample is insufficient.

Mariculture is crucial in environmental monitoring and safety assurance of marine environments. Certain mariculture areas are often partially or completely submerged in water, which causes the target signal to be extremely weak and difficult to detect. A method of target recognition and classification based on the convolutional neural network called semantic segmentation can fully consider the space spectrum and context semantic information. Therefore, this study proposes a target extraction method on the basis of multisource feature fusion, such as nNDWI and G/R ratio. In this work, the proposed recognition algorithm is verified under the conditions of uniform distribution of strong, weak, and extremely weak signals. Results show that the G/R feature is superior under the condition of uniform distribution of strong and weak signals. Its mean pixels accuracy is 2.32% higher than RGB (combination of red band, green band, and blue band), and its overall classification accuracy is 98.84%. Under the condition of extremely weak signal, the MPA of the multisource feature method based on the combination of G/R and nNDWI is 10.76% higher than RGB, and the overall classification accuracy is 82.02%. Under this condition, the G/R features highlight the target, and nNDWI suppresses noise. The proposed method can effectively extract the information of weak signal in the marine culture area and provide technical support for marine environmental monitoring and marine safety assurance.