Multi-scale Image Segmentation Method Research Articles

Integrating Sentinel-1/2 Data and Machine Learning to Map Cotton Fields in Northern Xinjiang, China

Timely and accurate information of cotton planting areas is essential for monitoring and managing cotton fields. However, there is no large-scale and high-resolution method suitable for mapping cotton fields, and the problems associated with low resolution and poor timeliness need to be solved. Here, we proposed a new framework for mapping cotton fields based on Sentinel-1/2 data for different phenological periods, random forest classifiers, and the multi-scale image segmentation method. A cotton field map for 2019 at a spatial resolution of 10 m was generated for northern Xinjiang, a dominant cotton planting region in China. The overall accuracy and kappa coefficient of the map were 0.932 and 0.813, respectively. The results showed that the boll opening stage was the best phenological phase for mapping cotton fields and the cotton fields was identified most accurately at the early boll opening stage, about 40 days before harvest. Additionally, Sentinel-1 and the red edge bands in Sentinel-2 are important for cotton field mapping, and there is great potential for the fusion of optical images and microwave images in crop mapping. This study provides an effective approach for high-resolution and high-accuracy cotton field mapping, which is vital for sustainable monitoring and management of cotton planting.

Context-Enabled Extraction of Large-Scale Urban Functional Zones from Very-High-Resolution Images: A Multiscale Segmentation Approach

Urban functional-zone (UFZ) analysis has been widely used in many applications, including urban environment evaluation, and urban planning and management. How to extract UFZs’ spatial units which delineates UFZs’ boundaries is fundamental to urban applications, but it is still unresolved. In this study, an automatic, context-enabled multiscale image segmentation method is proposed for extracting spatial units of UFZs from very-high-resolution satellite images. First, a window independent context feature is calculated to measure context information in the form of geographic nearest-neighbor distance from a pixel to different image classes. Second, a scale-adaptive approach is proposed to determine appropriate scales for each UFZ in terms of its context information and generate the initial UFZs. Finally, the graph cuts algorithm is improved to optimize the initial UFZs. Two datasets including WorldView-2 image in Beijing and GaoFen-2 image in Nanchang are used to evaluate the proposed method. The results indicate that the proposed method can generate better results from very-high-resolution satellite images than widely used approaches like image tiles and road blocks in representing UFZs. In addition, the proposed method outperforms existing methods in both segmentation quality and running time. Therefore, the proposed method appears to be promising and practical for segmenting large-scale UFZs.

Region-Line Association Constraints for High-Resolution Image Segmentation

Man-made objects, including buildings and roads, often feature straight boundaries. Specific shape features, including parallel boundaries (e.g., roads and buildings) and perpendicular corners (e.g., buildings), are strong structural clues for distinguishing man-made objects from natural objects. In this study, such observations are implemented in remote-sensing image segmentation. Several region-line association constraints are proposed, including the parallel straight line (PLSL) neighborhood, the perpendicular straight line (PPSL) neighborhood, and the PLSL zone. A region-merging process with these structural constraints is appended to the hard-boundary-constrained segmentation method (HBC-SEG), which is a multiscale image segmentation method. Results show that, compared with HBC-SEG, the proposed method presents a significant decrease in oversegmentation errors (measured by recall ratio r) and an insignificant increase in undersegmentation errors (measured by precision p). As a result, measure m 2 , which synthetically evaluates undersegmentation and oversegmentation errors, increases significantly. The improved method helps obtain complete objects, thus facilitating feature extraction and image classification for object-based image analysis.

다중스케일 노멀라이즈 컷을 이용한 영상분할

본 논문은 기존 그래프 컷 기반 영상분할의 성능은 유지하면서 연산속도가 빠른 영상분할 방법을 제안한다. 기존 그래프 컷 기반 영상분할은 높은 성능을 보이지만 고유쌍 연산으로 인해 분할 속도가 느리다는 단점을 지닌다. 이는 고유쌍 연산에서 영상 내 모든 화소 사이의 유사도를 고려하여 정방행렬을 만들기 때문이다. 그러므로 제안하는 방법은 영상을 여러 영역으로 분할하여 작은 크기의 정방행렬을 구성하고 이를 통해 고유쌍 연산 속도를 크게 향상시킨다. 본 논문에서는 대수적 다중 격자를 이용한 다중스케일 영상분할법을 제안하고 실험 결과를 통해 제안하는 방법이 기존 영상분할 방법보다 그 성능이 더 우수함을 보인다. This paper proposes a fast image segmentation method that gives high segmentation performance as graph-cut based methods. Graph-cut based image segmentation methods show high segmentation performance, however, the computational complexity is high to solve a computationally-intensive eigen-system. This is because solving eigen-system depends on the size of square matrix obtained from similarities between all pairs of pixels in the input image. Therefore, the proposed method uses the small-size square matrix, which is obtained from all the similarities among regions obtained by segmenting locally an image into several regions by graph-based method. Experimental results show that the proposed multi-scale image segmentation method using the algebraic multi-grid shows higher performance than existing methods.