Journal of Geo-information Science >
Waterbody Extraction from SAR Imagery based on Improved Speckle Reducing Anisotropic Diffusion and Maximum Between-Cluster Variance
Received date: 2018-10-18
Request revised date: 2019-03-01
Online published: 2019-06-15
Supported by
National Natural Science Foundation of China, No.41301479
Liaoning Province Natural Science Foundation, No.2015020090
Copyright
The use of remote sensing technology to obtain surface waterbody information is of great significance for water resource investigation, natural disaster assessment, watershed planning, and ecological environment monitoring. As a reliable data source for large-scale ground monitoring, SAR imaging has unique advantages that optical remote sensing systems of all-weather, all-weather, and wide coverage do not have, and has been widely used in waterbody mapping. However, due to the influence of speckle noise of SAR imagery, existing methods for waterbody mapping are difficult to extract the complex and fine natural water structures from SAR imagery quickly and accurately. In this paper, a new waterbody extraction method for SAR imagery based on improved speckle reducing anisotropic diffusion and maximum between-cluster variance was proposed. First, the SAR imagery were filtered by improved speckle reducing anisotropic diffusion. The iterative process was adaptively controlled by calculating the average structural similarity between imagery in the iterative filtering process, so that the fine edges and texture structure could be preserved simultaneously. Then, based on the criterion of maximum variance between classes, the threshold value was determined adaptively, and the binary segmentation of the filtered image was conducted. In the binarized segmentation result, connected foreground regions composed of the pixel points that have same intensity values and adjoin to each other spatially, are searched. In so doing, each connected region formed an identified block. By obtaining geometric parameters of these blocks, the false segmentation of the imagery was eliminated, and real waterbody areas were precisely identified based on the SAR imagery. To verify the accuracy of the proposed method, water boundaries extracted by this method were on manually drawn waterbody boundaries. Results of comparing the two methods show that they are pretty consistent with each other. Meanwhile, the results of our proposed method were compared with the results of three other kinds of water extraction algorithms commonly used for SAR imagery, in terms of the visual level, extraction accuracy, and running time. The running time of the proposed method meets the requirement of real-time application. The overlap degree of the extracted boundary in two pixels rating areas has reached 80%, which is obviously superior to other methods and the extraction results of the proposed method are also more significant in visual aspects such as boundary and detail information. The qualitative and quantitative evaluation shows the superiority of our proposed method.
LI Yu , YANG Yun , ZHAO Quanhua . Waterbody Extraction from SAR Imagery based on Improved Speckle Reducing Anisotropic Diffusion and Maximum Between-Cluster Variance[J]. Journal of Geo-information Science, 2019 , 21(6) : 907 -917 . DOI: 10.12082/dqxxkx.2019.180525
Fig. 1 Original intensity maps of the SAR imagery图1 原始SAR强度图像 |
Fig. 2 Flowchart of the method of this paper图2 本文方法的流程 |
Fig. 3 SRAD edge detection diagram图3 SRAD边缘检测示意 |
Fig. 4 SRAD first-order neighborhood Diffusion diagram图4 SRAD一阶邻域扩散示意 |
Fig. 5 Flowchart of the pseudo waterbody removal algorithm based on connected-domain calibration图5 基于连通域标定的伪水体去除算法流程 |
Fig. 6 SARD filtering results图6 SARD滤波结果 |
Fig. 7 Extraction results and qualitative evaluation of this paper图7 本文方法提取结果和定性评价 |
Fig. 8 Pixel buffer boundary overlay analysis图8 像元缓冲区边界叠加分析示意 |
Tab. 1 Extraction accuracy of the 5 SAR images in the case of different radius rating zones (%)表1 在不同半径评级区的情况下5幅SAR图像的提取精度对比 |
重叠 | 1个像元 | 2个像元 | 3个像元 | 4个像元 | 5个像元 | |
---|---|---|---|---|---|---|
图1(a) | 21.05 | 68.90 | 90.21 | 96.76 | 99.23 | 99.54 |
图1(b) | 19.13 | 60.11 | 88.56 | 93.62 | 96.23 | 97.04 |
图1(c) | 18.39 | 60.29 | 82.16 | 90.32 | 92.90 | 94.12 |
图1(d) | 8.27 | 59.84 | 81.66 | 92.48 | 93.77 | 95.93 |
图1(e) | 20.29 | 69.14 | 91.01 | 92.81 | 95.60 | 97.05 |
Fig. 9 Comparison of extraction results of Otsu, level set and reference [17] method图9 Otsu、水平集和文献[17]方法提取结果比较 |
Fig. 10 Otsu, level set and literature [17] method contour overlay results图10 Otsu、水平集和文献[17]方法轮廓线叠加结果 |
Tab. 2 Extraction accuracy of the 5 SAR images with different radius rating zones by the compared algorithms (%)表2 对比算法在不同半径评级区的情况下5幅SAR图像的提取精度对比 |
图像 | 方法 | 重叠 | 1个像元 | 2个像元 | 3个像元 | 4个像元 | 5个像元 |
---|---|---|---|---|---|---|---|
图1(a) | Otsu | 3.88 | 12.57 | 18.82 | 23.60 | 28.04 | 31.71 |
水平集 | 11.42 | 37.24 | 53.89 | 62.94 | 68.86 | 72.36 | |
文献[17] | 12.58 | 42.52 | 59.64 | 64.87 | 67.45 | 69.31 | |
图1(b) | Otsu | 4.16 | 13.25 | 19.72 | 23.87 | 27.11 | 30.06 |
水平集 | 9.84 | 30.29 | 43.03 | 49.27 | 53.51 | 56.75 | |
文献[17] | 15.44 | 46.78 | 62.08 | 66.28 | 67.63 | 69.36 | |
图1(c) | Otsu | 2.79 | 8.72 | 12.79 | 16.52 | 17.86 | 19.37 |
水平集 | 4.85 | 10.92 | 17.00 | 21.56 | 22.71 | 23.31 | |
文献[17] | 9.27 | 26.79 | 34.08 | 37.23 | 38.48 | 39.13 | |
图1(d) | Otsu | 1.19 | 5.65 | 9.43 | 11.12 | 12.50 | 13.37 |
水平集 | 4.96 | 11.47 | 18.30 | 22.57 | 24.96 | 26.06 | |
文献[17] | 10.96 | 23.47 | 31.30 | 36.57 | 39.96 | 41.06 | |
图1(e) | Otsu | 5.75 | 18.30 | 27.91 | 34.50 | 39.36 | 43.48 |
水平集 | 7.71 | 24.89 | 37.83 | 46.21 | 51.96 | 56.40 | |
文献[17] | 15.41 | 49.18 | 69.24 | 76.46 | 78.39 | 79.30 |
Tab. 3 Comparison of extraction time between the compared algorithms and the method of this paper (s)表3 对比算法与本文方法的水体提取时间对比 |
方法 | 图1(a) | 图1(b) | 图1(c) | 图1(d) | 图1(e) |
---|---|---|---|---|---|
Otsu | 1.205 | 1.227 | 4.380 | 15.608 | 4.633 |
水平集 | 224.873 | 216.374 | 1007.716 | 5068.251 | 1179.581 |
文献[17] | 9.527 | 8.371 | 56.153 | 223.176 | 78.113 |
本文方法 | 4.382 | 4.178 | 15.322 | 29.644 | 16.013 |
The authors have declared that no competing interests exist.
[1] |
|
[2] |
[
|
[3] |
[
|
[4] |
[
|
[5] |
[
|
[6] |
[
|
[7] |
|
[8] |
|
[9] |
[
|
[10] |
[
|
[11] |
[
|
[12] |
[
|
[13] |
[
|
[14] |
|
[15] |
[
|
[16] |
[
|
[17] |
[
|
[18] |
[
|
[19] |
|
[20] |
|
[21] |
|
[22] |
|
[23] |
http://rs.ceode.ac.cn/satelliteintroduce/introRadarsats.jsp# RADARSAT-1
|
[24] |
http://cloud.satimage.cn/data_intro
|
[25] |
http://www.cresda.com/CN/sjfw/zxsj/index.shtml
|
[26] |
|
[27] |
|
[28] |
|
/
〈 | 〉 |