CN114612773A - Efficient sea ice motion extraction method and system suitable for SAR and optical images - Google Patents
Efficient sea ice motion extraction method and system suitable for SAR and optical images Download PDFInfo
- Publication number
- CN114612773A CN114612773A CN202210178701.9A CN202210178701A CN114612773A CN 114612773 A CN114612773 A CN 114612773A CN 202210178701 A CN202210178701 A CN 202210178701A CN 114612773 A CN114612773 A CN 114612773A
- Authority
- CN
- China
- Prior art keywords
- sea ice
- image
- motion vector
- sar
- ice motion
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 230000033001 locomotion Effects 0.000 title claims abstract description 125
- 238000000605 extraction Methods 0.000 title claims abstract description 57
- 230000003287 optical effect Effects 0.000 title claims abstract description 36
- 238000000034 method Methods 0.000 claims abstract description 79
- 239000013598 vector Substances 0.000 claims abstract description 70
- 238000001914 filtration Methods 0.000 claims abstract description 25
- 238000007781 pre-processing Methods 0.000 claims abstract description 8
- 238000007670 refining Methods 0.000 claims abstract 2
- 238000001514 detection method Methods 0.000 claims description 8
- 230000011218 segmentation Effects 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 7
- 238000012634 optical imaging Methods 0.000 claims description 6
- 230000006870 function Effects 0.000 claims description 4
- 230000003044 adaptive effect Effects 0.000 claims 2
- 230000005764 inhibitory process Effects 0.000 claims 2
- 230000000694 effects Effects 0.000 claims 1
- 230000007547 defect Effects 0.000 abstract description 3
- 230000008569 process Effects 0.000 description 7
- 230000001629 suppression Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 238000007792 addition Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000009792 diffusion process Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 238000005553 drilling Methods 0.000 description 1
- 230000008278 dynamic mechanism Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000003344 environmental pollutant Substances 0.000 description 1
- 239000013505 freshwater Substances 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 231100000719 pollutant Toxicity 0.000 description 1
Images
Landscapes
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
Description
技术领域technical field
本发明属于冰冻圈海冰遥感领域，特别涉及一种适用于SAR和光学影像的高效海冰运动提取方法及系统。The invention belongs to the field of cryosphere sea ice remote sensing, and particularly relates to an efficient sea ice motion extraction method and system suitable for SAR and optical images.
背景技术Background technique
海冰运动是海冰在大气、洋流、地转偏向力等多种因素作用下的结果。海冰和气候之间存在很强的耦合，对海冰运动进行研究，有助于理解和预测未来极区甚至全球的气候变化。海冰运动是北冰洋海冰面积变化的重要动力机制，对北冰洋海冰的物质平衡起着重要作用；海冰运动也是海洋中潜热和淡水的主要输运方式，影响海洋盐度和局部热交换。掌握海冰的运动状态，可以有效地阻止海冰对航行船舶和石油钻井平台的危害。生态环境方面，海冰运动造成的污染物扩散已经成为不可忽视的问题。Sea ice movement is the result of sea ice under the action of various factors such as atmosphere, ocean currents, and geostrophic deflection forces. There is a strong coupling between sea ice and climate, and the study of sea ice movement is helpful to understand and predict future climate changes in polar regions and even the world. Sea ice movement is an important dynamic mechanism for the change of sea ice area in the Arctic Ocean, and plays an important role in the material balance of sea ice in the Arctic Ocean; sea ice movement is also the main transport mode of latent heat and fresh water in the ocean, affecting ocean salinity and local heat exchange. Mastering the movement state of sea ice can effectively prevent the harm of sea ice to ships and oil drilling platforms. In terms of ecological environment, the diffusion of pollutants caused by sea ice movement has become a problem that cannot be ignored.
卫星遥感影像是获得海冰运动的重要数据源，从遥感影像中提取海冰运动的方法，可以分为三类，分别是：(1)差分方法；(2)模板匹配；(3)特征追踪。第一种类型的方法以光流法为代表，能够获得稠密的海冰运动场，但是该方法受制于光流法的缺点，不适合对有较大位移的海冰运动进行提取。第二种方法采用度量影像块的相似性进行海冰运动矢量提取，此类方法中有最大互相关法(MCC)、相位相关法(PC)，其中MCC在高分辨率和低分辨率海冰运动提取中使用最为广泛。模板匹配方法实现简单、鲁棒性好，但是此种方法不能检测海冰的旋转运动，并且存在计算效率低的问题。第三种方法通过特征提取和匹配进行海冰运动提取，该方法能够有效检测海冰的旋转运动，并且对海冰的形变具有较好的鲁棒性，但是获得的海冰运动矢量分布稀疏且不均匀，对海冰运动数据的进一步使用造成影响。Satellite remote sensing images are an important data source for obtaining sea ice movement. The methods of extracting sea ice movement from remote sensing images can be divided into three categories: (1) difference method; (2) template matching; (3) feature tracking . The first type of method is represented by the optical flow method, which can obtain a dense sea ice motion field, but this method is subject to the shortcomings of the optical flow method and is not suitable for the extraction of sea ice motion with large displacement. The second method uses the similarity of the metric image blocks to extract the motion vector of sea ice. There are maximum cross-correlation method (MCC) and phase correlation method (PC) in such methods. MCC is used in high-resolution and low-resolution sea ice It is the most widely used in motion extraction. The template matching method is simple to implement and has good robustness, but this method cannot detect the rotational motion of sea ice and has the problem of low computational efficiency. The third method extracts sea ice motion through feature extraction and matching. This method can effectively detect the rotational motion of sea ice and has good robustness to the deformation of sea ice, but the obtained sea ice motion vector is sparsely distributed and Uneven, with implications for further use of sea ice movement data.
另外，一些结合特征追踪和模板匹配海冰运动提取方法优点的混合方法被提出来，但是混合方法仅是在利用单一数据源的基础上提出的，普适性差；混合方法在特征点匹配阶段采用效率低下的暴力匹配方法，随着特征点数量的增加，暴力匹配方法的效率呈现指数下降的趋势。In addition, some hybrid methods combining the advantages of feature tracking and template matching sea ice motion extraction methods have been proposed, but the hybrid methods are only proposed on the basis of using a single data source, and the generalizability is poor; the hybrid methods are used in the feature point matching stage. Inefficient brute force matching method, with the increase of the number of feature points, the efficiency of brute force matching method shows an exponential decreasing trend.
发明内容SUMMARY OF THE INVENTION
为了解决目前海冰运动提取方法普适性差以及特征点匹配效率低的问题，本发明提供一种适用于SAR和光学影像的高效海冰运动提取方案。In order to solve the problems of poor universality and low feature point matching efficiency of current sea ice motion extraction methods, the present invention provides an efficient sea ice motion extraction scheme suitable for SAR and optical images.
为了实现上述目的，本发明提出的技术方案为一种适用于SAR和光学影像的高效海冰运动提取方法，包括如下步骤：In order to achieve the above purpose, the technical solution proposed by the present invention is an efficient sea ice motion extraction method suitable for SAR and optical images, including the following steps:
步骤1，影像预处理，包括将SAR影像的后向散射系数通过线性拉伸，转换为灰度影像，对光学影像各通道像素值进行加权平均获得单通道灰度影像；Step 1, image preprocessing, including converting the backscattering coefficient of the SAR image into a grayscale image through linear stretching, and performing a weighted average of the pixel values of each channel of the optical image to obtain a single-channel grayscale image;
步骤2，读入两幅预处理之后的影像，两幅影像具有重叠区域，且获取时间小于预设的时间段，利用改进ORB算法分别对两幅影像进行特征点提取，实现方式为，构建影像金字塔，确定每层影像需要提取的特征点数量，对每层影像进行规则格网划分，对每层影像每个格网内的影像块采用加速分段检测特征点提取方式进行特征点提取，再采用四叉树非极大抑制方法均匀选取每层影像指定数量的特征点；Step 2: Read in two preprocessed images, the two images have overlapping areas, and the acquisition time is less than a preset time period, and the improved ORB algorithm is used to extract feature points from the two images respectively. The implementation method is: constructing an image Pyramid, determine the number of feature points that need to be extracted for each layer of image, divide each layer of image into regular grid, use accelerated segmentation detection feature point extraction method to extract feature points for each image block in each grid of each layer of image, and then The quadtree non-maximum suppression method is used to uniformly select a specified number of feature points in each image layer;
步骤3，采用基于地理网格的匹配方法，对两幅影像的特征点进行匹配；所述基于地理网格的匹配方法，实现方式为按照地理网格对影像上的特征点进行分组，计算每个特征点所在网格的行列号，将行列号相同的特征点分为一组；按照海冰的最大漂移速度计算匹配半径，对两幅影像特征点进行匹配；Step 3, using a matching method based on geographic grids to match the feature points of the two images; the matching method based on geographic grids is implemented by grouping the feature points on the images according to the geographic grids, and calculating each The row and column numbers of the grid where the feature points are located, and the feature points with the same row and column numbers are grouped into a group; the matching radius is calculated according to the maximum drift speed of the sea ice, and the feature points of the two images are matched;
步骤4，根据匹配点对的位置以及两幅影像的时间差，计算初始海冰运动矢量；Step 4: Calculate the initial sea ice motion vector according to the position of the matched point pair and the time difference between the two images;
步骤5，根据初始海冰运动矢量，按照指定空间分辨率规则地内插海冰运动矢量；Step 5, according to the initial sea ice motion vector, regularly interpolate the sea ice motion vector according to the specified spatial resolution;
步骤6，针对内插所得海冰运动矢量和真实的海冰运动矢量存在的偏差，按照最大互相关方法进行精化；Step 6, for the deviation between the sea ice motion vector obtained by interpolation and the real sea ice motion vector, refine according to the maximum cross-correlation method;
步骤7，采用局部一致性滤波对精化后的海冰运动矢量场进行滤波，剔除错误的海冰运动矢量，获得海冰运动提取结果。Step 7: Use local consistency filtering to filter the refined sea ice motion vector field, remove wrong sea ice motion vectors, and obtain sea ice motion extraction results.
而且，步骤2中采用改进ORB算法进行特征点提取时，每层影像每个格网内的影像块采用加速分段检测特征点提取方法进行特征点提取，包括首先采用较大阈值对格网内影像块进行特征点提取，如果没有提取到特征点则采用较小的阈值重新提取特征点。Moreover, when the improved ORB algorithm is used to extract feature points in step 2, the image blocks in each grid of each layer of image are extracted by using the accelerated segmentation detection feature point extraction method, including first using a larger threshold to extract the feature points in the grid. Feature point extraction is performed on the image block, and if no feature point is extracted, the feature point is re-extracted with a smaller threshold.
而且，步骤3中采用基于地理网格的特征点匹配方法中，按照海冰的最大漂移速度计算匹配半径实现如下，Moreover, in the feature point matching method based on geographic grid adopted in step 3, the matching radius is calculated according to the maximum drift speed of sea ice and is implemented as follows:
其中，smax为海冰的最大漂移速度，dt为两影像之间的时间间隔，gsd为地理网格分辨率，ceil为向上取整函数。where s max is the maximum drift velocity of sea ice, d t is the time interval between two images, gsd is the geographic grid resolution, and ceil is the round-up function.
而且，步骤7中采用局部一致性滤波方法对海冰运动矢量场进行滤波；局部一致性滤波考虑了运动矢量和周围运动矢量之间的差异性，对差异性大小进行判断，剔除差异性较大的运动矢量；通过对光学影像提取的海冰运动场进行滤波测试，局部一致性滤波取得了较好的效果。Moreover, in step 7, the local consistency filtering method is used to filter the sea ice motion vector field; the local consistency filtering considers the difference between the motion vector and the surrounding motion vector, judges the size of the difference, and eliminates the large difference Through the filtering test of the sea ice sports field extracted from the optical image, the local consistency filtering has achieved good results.
本发明还提供一种适用于SAR和光学影像的高效海冰运动提取系统，用于实现如上所述的一种适用于SAR和光学影像的高效海冰运动提取方法。The present invention also provides an efficient sea ice motion extraction system suitable for SAR and optical images, which is used to realize the above-mentioned efficient sea ice motion extraction method suitable for SAR and optical images.
而且，包括以下模块，Also, the following modules are included,
第一模块，用于影像预处理，包括将SAR影像的后向散射系数通过线性拉伸，转换为灰度影像，对光学影像各通道像素值进行加权平均获得单通道灰度影像；The first module is used for image preprocessing, including converting the backscatter coefficient of the SAR image into a grayscale image through linear stretching, and performing a weighted average of the pixel values of each channel of the optical image to obtain a single-channel grayscale image;
第二模块，用于读入两幅预处理之后的影像，两幅影像具有重叠区域，且获取时间小于预设的时间段，利用改进ORB算法分别对两幅影像进行特征点提取，实现方式为，构建影像金字塔，确定每层影像需要提取的特征点数量，对每层影像进行规则格网划分，对每层影像每个格网内的影像块采用加速分段检测特征点提取方式进行特征点提取，再采用四叉树非极大抑制方法均匀选取每层影像指定数量的特征点；The second module is used to read in two preprocessed images. The two images have overlapping areas and the acquisition time is less than the preset time period. The improved ORB algorithm is used to extract feature points from the two images respectively. The implementation method is as follows: , build an image pyramid, determine the number of feature points that need to be extracted for each layer of image, divide each layer of image into regular grid, and use accelerated segmentation detection feature point extraction method to extract feature points for each image block in each grid of each layer of image. Extraction, and then use the quadtree non-maximum suppression method to uniformly select a specified number of feature points in each layer of image;
第三模块，用于采用基于地理网格的匹配方法，对两幅影像的特征点进行匹配；所述基于地理网格的匹配方法，实现方式为按照地理网格对影像上的特征点进行分组，计算每个特征点所在网格的行列号，将行列号相同的特征点分为一组；按照海冰的最大漂移速度计算匹配半径，对两幅影像特征点进行匹配；The third module is used to match the feature points of the two images by using the matching method based on geographic grid; the matching method based on geographic grid is realized by grouping the feature points on the images according to the geographic grid , calculate the row and column number of the grid where each feature point is located, and group the feature points with the same row and column number into a group; calculate the matching radius according to the maximum drift speed of the sea ice, and match the two image feature points;
第四模块，用于根据匹配点对的位置以及两幅影像的时间差，计算初始海冰运动矢量；The fourth module is used to calculate the initial sea ice motion vector according to the position of the matched point pair and the time difference between the two images;
第五模块，用于根据初始海冰运动矢量，按照指定空间分辨率规则地内插海冰运动矢量；The fifth module is used to regularly interpolate the sea ice motion vector according to the specified spatial resolution according to the initial sea ice motion vector;
第六模块，用于针对内插所得海冰运动矢量和真实的海冰运动矢量存在的偏差，按照最大互相关方法进行精化；The sixth module is used to refine the deviation between the sea ice motion vector obtained by interpolation and the real sea ice motion vector according to the maximum cross-correlation method;
第七模块，用于采用局部一致性滤波对精化后的海冰运动矢量场进行滤波，剔除错误的海冰运动矢量，获得海冰运动提取结果。The seventh module is used to filter the refined sea ice motion vector field by using local consistency filtering, remove the wrong sea ice motion vector, and obtain the sea ice motion extraction result.
或者，包括处理器和存储器，存储器用于存储程序指令，处理器用于调用存储器中的存储指令执行如上所述的一种适用于SAR和光学影像的高效海冰运动提取方法。Alternatively, it includes a processor and a memory, the memory is used to store program instructions, and the processor is used to call the stored instructions in the memory to execute the above-mentioned efficient sea ice motion extraction method suitable for SAR and optical images.
或者，包括可读存储介质，所述可读存储介质上存储有计算机程序，所述计算机程序执行时，实现如上所述的一种适用于SAR和光学影像的高效海冰运动提取方法。Alternatively, a readable storage medium is included, a computer program is stored on the readable storage medium, and when the computer program is executed, an efficient sea ice motion extraction method suitable for SAR and optical images as described above is implemented.
与现有技术相比，本发明具有以下优点：Compared with the prior art, the present invention has the following advantages:
(1)本发明能够利用SAR和光学影像进行海冰运动提取，对数据源类型具有普适性，为利用多源数据生产长时序的海冰运动产品提供可能。(1) The present invention can use SAR and optical images to extract sea ice movement, has universality to data source types, and provides the possibility to use multi-source data to produce long-term sea ice movement products.
(2)本发明采用基于地理网格的匹配方法，提高了海冰运动提取的效率，为大规模、批量的海冰运动提取提供可能。(2) The present invention adopts the matching method based on geographic grid, which improves the efficiency of sea ice movement extraction, and provides possibility for large-scale and batch sea ice movement extraction.
本发明方案实施简单方便，实用性强，解决了相关技术存在的实用性低及实际应用不便的问题，能够提高用户体验，具有重要的市场价值。The solution of the invention is simple and convenient to implement and has strong practicability, solves the problems of low practicability and inconvenient practical application of the related technologies, can improve user experience, and has important market value.
附图说明Description of drawings
图1为本发明实施例改进ORB算法提取特征点流程图。FIG. 1 is a flowchart of an improved ORB algorithm for extracting feature points according to an embodiment of the present invention.
图2为本发明实施例基于地理网格的匹配方法示意图。FIG. 2 is a schematic diagram of a matching method based on a geographic grid according to an embodiment of the present invention.
图3为本发明实施例一种适用于SAR和光学影像的高效海冰运动提取方法示意图。FIG. 3 is a schematic diagram of an efficient sea ice motion extraction method suitable for SAR and optical images according to an embodiment of the present invention.
具体实施方式Detailed ways
以下结合附图和实施例具体说明本发明的技术方案。The technical solutions of the present invention will be specifically described below with reference to the accompanying drawings and embodiments.
本发明公开了一种适用于SAR和光学影像的高效海冰运动提取方案，克服了以往海冰运动提取方法仅能利用单一数据源的缺陷；充分利用了影像的地理位置信息，提升了海冰运动提取的效率。本发明方法的基本流程是，对SAR(例如Sentinel-1SAR、EnvisatASAR、ALOS PALSAR、ALOS-2PALSAR-2、GF-3SAR)和光学影像(MODIS)进行预处理，获得灰度级为0-255的灰度影像；采用改进的ORB算法对灰度影像进行特征点提取；采用基于地理网格的匹配方法对特征点进行匹配获得初始海冰运动矢量；利用初始海冰运动矢量按照指定空间分辨率内插海冰运动矢量；利用最大互相关对内插的海冰运动矢量进行精化；对精化后的海冰运动矢量场进行滤波，剔除错误的矢量。The invention discloses an efficient sea ice movement extraction scheme suitable for SAR and optical images, overcomes the defect that the previous sea ice movement extraction method can only use a single data source; fully utilizes the geographic location information of the image, and improves the sea ice movement Efficiency of motion extraction. The basic flow of the method of the present invention is to preprocess SAR (for example, Sentinel-1SAR, EnvisatASAR, ALOS PALSAR, ALOS-2PALSAR-2, GF-3SAR) and optical image (MODIS) to obtain a gray level of 0-255. Grayscale image; use the improved ORB algorithm to extract feature points from the grayscale image; use the matching method based on geographic grid to match the feature points to obtain the initial sea ice motion vector; use the initial sea ice motion vector according to the specified spatial resolution. Interpolate sea ice motion vector; use maximum cross-correlation to refine the interpolated sea ice motion vector; filter the refined sea ice motion vector field to remove wrong vectors.
如图3所示，本发明实施例提供的一种适用于SAR和光学影像的高效海冰运动提取方法，具体包括以下步骤：As shown in FIG. 3 , an efficient sea ice motion extraction method suitable for SAR and optical images provided by an embodiment of the present invention specifically includes the following steps:
1.影像预处理，将SAR影像的后向散射系数通过线性拉伸，转换为灰度值在0-255的灰度影像；对光学影像各通道像素值进行加权平均获得灰度值在0-255的单通道灰度影像：实施例中根据式(1)将SAR影像的后向散射系数σ0转换为0-255的灰度值i；根据式(2)将光学影像的R、G、B值转换为单一灰度值i。1. Image preprocessing, convert the backscattering coefficient of the SAR image into a grayscale image with a grayscale value of 0-255 through linear stretching; the pixel value of each channel of the optical image is weighted and averaged to obtain a grayscale value of 0-255. 255 single-channel grayscale image: in the embodiment, the backscattering coefficient σ0 of the SAR image is converted into a grayscale value i of 0-255 according to formula (1); The B value is converted to a single gray value i.
i＝0.3·R+0.59·G+0.11·B (2)i=0.3·R+0.59·G+0.11·B (2)
其中，
2.读入两幅预处理之后的影像，两幅影像要求具有一定的重叠区域，且获取时间小于预设的时间段，具体实施时可以预设具体的重叠比例，3天；采用改进ORB算法分别对两幅影像进行特征点提取，提取的特征点集合分别标记为Kp1、Kp2：实施例中对预处理后的两幅具有重叠区域且获取时间间隔小于3天的影像，并针对本发明的海冰运动提取需求提出利用改进的ORB算法进行特征点提取。相比原始ORB算法，改进ORB算法主要对特征点的筛选策略进行了改进，其流程如图1所示，详细过程为：按照指定的比例因子q，构建具有L层结构的影像金字塔；根据指定的特征点数量N，按照公式(3)计算第k层影像需要提取的特征点数量Nk；对每层影像进行规则格网划分，格网的大小为30pixel×30pixel；对每层影像每个格网内的影像块采用加速分段检测特征点提取(oFAST)算法进行特征点提取，即首先采用较大阈值(优选为15)对格网内影像块进行特征点提取，如果没有提取到特征点则采用较小的阈值(优选为5)重新提取特征点；采用四叉树非极大抑制方法均匀选取每层影像指定数量的特征点。改进ORB算法采用双阈值，保证在影像的弱纹理区域也能提取到一定数量的特征点；之后采用四叉树非极大抑制方法，均匀选择出目标数量的特征点，保证特征点的数量不至于过多，同时保证特征点在影像上分布均匀。改进ORB算法克服了ORB算法提取的特征点过于集中在海岸线、冰脊、冰间水道的缺陷。四叉树非极大抑制方法具体实现为现有技术，本发明不予赘述。2. Read in two preprocessed images. The two images must have a certain overlap area, and the acquisition time is less than the preset time period. The specific overlap ratio can be preset during the specific implementation, 3 days; the improved ORB algorithm is adopted Feature point extraction is performed on the two images respectively, and the extracted feature point sets are marked as Kp1 and Kp2 respectively. It is proposed to use the improved ORB algorithm to extract feature points. Compared with the original ORB algorithm, the improved ORB algorithm mainly improves the selection strategy of feature points. The process is shown in Figure 1. The detailed process is: according to the specified scale factor q, build an image pyramid with an L-layer structure; According to formula (3), calculate the number of feature points N k that needs to be extracted from the image of the kth layer; divide the image of each layer into a regular grid, and the size of the grid is 30pixel×30pixel; The image blocks in the grid are extracted by the accelerated segmentation detection feature point extraction (oFAST) algorithm, that is, a larger threshold (preferably 15) is used to extract the feature points of the image blocks in the grid. For points, a smaller threshold (preferably 5) is used to re-extract feature points; a quadtree non-maximum suppression method is used to uniformly select a specified number of feature points in each image layer. The improved ORB algorithm uses double thresholds to ensure that a certain number of feature points can also be extracted in the weak texture area of the image; then the quadtree non-maximum suppression method is used to uniformly select the target number of feature points to ensure that the number of feature points does not vary. As for too much, at the same time ensure that the feature points are evenly distributed on the image. The improved ORB algorithm overcomes the defect that the feature points extracted by the ORB algorithm are too concentrated on coastlines, ice ridges, and interglacial waterways. The quadtree non-maximum suppression method is specifically implemented in the prior art, and will not be described in detail in the present invention.
3.采用基于地理网格的匹配方法，对提取的特征点进行匹配：本发明提出的基于地理网格的匹配方法如图2所示，详细过程为：按照地理网格EASE Grid 2.0对影像上的特征点进行分组，计算每个特征点所在网格的行列号，将行列号相同的特征点分为一组，记影像1(image1)上特征点分组的结果为[G11,G12,…,G1n]，影像2(image2)上特征点分组结果为[G21,G22,…,G2m]，其中n、m分别表示影像1上特征点分为n组，影像2上特征点分为m组；按照公式(4)计算匹配半径R，其中smax为海冰的最大漂移速度0.5m/s，dt为两影像之间的时间间隔，gsd为地理网格分辨率，ceil为向上取整函数；对影像1中第i组特征点G1i进行匹配，记G1i在EASE Grid 2.0中行、列索引分别为ri、ci，在影像2的分组结果中取出行索引在[ri-R,ri+R]范围、列索引在[ci-R,ci+R]范围内的特征点组，合并这些特征点组为G2i，将G1i与G2i中的特征点按照汉明距离进行匹配，获得匹配点组Mi。基于地理网格的特征点匹配方法，利用了特征点的地理信息，降低了特征点匹配的搜索范围，相比于传统的暴力匹配方法，效率提升了8～10倍。3. Use the matching method based on geographic grid to match the extracted feature points: The matching method based on geographic grid proposed by the present invention is shown in Figure 2, and the detailed process is: according to geographic grid EASE Grid 2.0 The feature points are grouped, the row and column numbers of the grid where each feature point is located are calculated, the feature points with the same row and column numbers are grouped into a group, and the result of the feature point grouping on image 1 (image1) is [G 11 , G 12 , . _ _ _ The points are divided into m groups; the matching radius R is calculated according to formula (4), where s max is the maximum drift speed of sea ice 0.5m/s, d t is the time interval between two images, gsd is the geographic grid resolution, ceil is an upward rounding function; match the i-th group of feature points G 1i in image 1, and denote that G 1i in EASE Grid 2.0 has the row and column indices as ri and c i respectively , and take out the row index from the grouping result of image 2 Feature point groups in the range of [r i -R,r i +R] and column indices in the range of [ ci -R, ci +R], merge these feature point groups into G 2i , and combine G 1i with G 2i The feature points in are matched according to the Hamming distance, and the matched point group M i is obtained. The feature point matching method based on geographic grid utilizes the geographic information of feature points to reduce the search range of feature point matching. Compared with the traditional brute force matching method, the efficiency is improved by 8 to 10 times.
4.根据影像的地理信息，将匹配点对的像素坐标转换为地理坐标；根据匹配点对的地理坐标(px,py)和(px′,p′y)以及影像对时间差Δt，按照公式(5)计算初始海冰运动矢量(vx,vy)。4. According to the geographic information of the image, convert the pixel coordinates of the matching point pair into geographic coordinates; Calculate the initial sea ice motion vector (v x , v y ) according to formula (5).
5.利用特征追踪获得的初始海冰运动矢量，预估格网化分布的运动矢量，详细过程为：根据指定空间分辨率生成规则格网，规则格网点所在位置的海冰运动矢量通过线性插值的方法获取。5. Use the initial sea ice motion vector obtained by feature tracking to estimate the motion vector of grid distribution. The detailed process is: generate a regular grid according to the specified spatial resolution, and the sea ice motion vector at the location of the regular grid point is linearly interpolated method to obtain.
6.采用最大互相关方法精化插值获得的规则格网点处的海冰运动矢量。最大互相关方法中相关系数ρ采用公式(6)进行计算。精化的过程即是在影像2的搜索区域中定位与影像1中模板相关系数最大的位置。6. The maximum cross-correlation method is used to refine the motion vector of sea ice at the regular grid points obtained by interpolation. In the maximum cross-correlation method, the correlation coefficient ρ is calculated by formula (6). The process of refinement is to locate the position with the largest correlation coefficient with the template in image 1 in the search area of image 2.
式中aij表示影像1中匹配模板的第i行、第j列的像素值，bij表示影像2中搜索区域子集的第i行、第j列的像素值，
7.采用局部一致性流场滤波方法，剔除海冰运动结果中的错误矢量，本发明提出的局部一致性流场滤波方法详细过程为：首先对投影坐标系下运动矢量的x方向分量进行局部一致性滤波，即计算海冰运动矢量x方向的速度大小与其周围相邻矢量x方向速度大小平均值之间的绝对差值，如果差值大于阈值，则剔除矢量；然后对投影坐标系下运动矢量的y方向分量进行局部一致性滤波，方法同x方向分量的局部一致性滤波；最后剔除周围矢量个数小于预设阈值(优选建议为4个)的海冰运动矢量。局部一致性滤波考虑了运动矢量和周围运动矢量之间的差异性，对差异性大小进行判断，剔除差异性较大的运动矢量。通过对光学影像提取的海冰运动场进行滤波测试，局部一致性滤波取得了较好的效果。7. The locally consistent flow field filtering method is adopted to eliminate the error vector in the result of sea ice motion. The detailed process of the locally consistent flow field filtering method proposed by the present invention is as follows: first, perform a localization on the x-direction component of the motion vector in the projected coordinate system. Consistency filtering, that is, calculating the absolute difference between the speed of the sea ice motion vector in the x direction and the average value of the speed of the surrounding adjacent vectors in the x direction. If the difference is greater than the threshold, the vector is eliminated; The y-direction component of the vector is subjected to local consistency filtering, and the method is the same as the local consistency filtering of the x-direction component; finally, the sea ice motion vectors whose number of surrounding vectors is less than the preset threshold (preferably 4 are recommended) are eliminated. The local consistency filter considers the difference between the motion vector and the surrounding motion vector, judges the size of the difference, and removes the motion vector with a large difference. Through the filtering test of the sea ice sports field extracted from the optical image, the local consistency filtering has achieved good results.
具体实施时，本发明技术方案提出的方法可由本领域技术人员采用计算机软件技术实现自动运行流程，实现方法的系统装置例如存储本发明技术方案相应计算机程序的计算机可读存储介质以及包括运行相应计算机程序的计算机设备，也应当在本发明的保护范围内。During specific implementation, the method proposed by the technical solution of the present invention can be realized by those skilled in the art by using computer software technology to realize the automatic running process. The system device for implementing the method is, for example, a computer-readable storage medium storing a computer program corresponding to the technical solution of the present invention, and a computer that runs the corresponding computer program. The computer equipment of the program should also be within the protection scope of the present invention.
在一些可能的实施例中，提供一种适用于SAR和光学影像的高效海冰运动提取系统，包括以下模块，In some possible embodiments, an efficient sea ice motion extraction system suitable for SAR and optical imaging is provided, including the following modules:
第一模块，用于影像预处理，包括将SAR影像的后向散射系数通过线性拉伸，转换为灰度影像，对光学影像各通道像素值进行加权平均获得单通道灰度影像；The first module is used for image preprocessing, including converting the backscatter coefficient of the SAR image into a grayscale image through linear stretching, and performing a weighted average of the pixel values of each channel of the optical image to obtain a single-channel grayscale image;
第二模块，用于读入两幅预处理之后的影像，两幅影像具有重叠区域，且获取时间小于预设的时间段，利用改进ORB算法分别对两幅影像进行特征点提取，实现方式为，构建影像金字塔，确定每层影像需要提取的特征点数量，对每层影像进行规则格网划分，对每层影像每个格网内的影像块采用加速分段检测特征点提取方式进行特征点提取，再采用四叉树非极大抑制方法均匀选取每层影像指定数量的特征点；The second module is used to read in two preprocessed images. The two images have overlapping areas and the acquisition time is less than the preset time period. The improved ORB algorithm is used to extract feature points from the two images respectively. The implementation method is as follows: , build an image pyramid, determine the number of feature points that need to be extracted for each layer of image, divide each layer of image into regular grid, and use accelerated segmentation detection feature point extraction method to extract feature points for each image block in each grid of each layer of image. Extraction, and then use the quadtree non-maximum suppression method to uniformly select a specified number of feature points in each layer of image;
第三模块，用于采用基于地理网格的匹配方法，对两幅影像的特征点进行匹配；所述基于地理网格的匹配方法，实现方式为按照地理网格对影像上的特征点进行分组，计算每个特征点所在网格的行列号，将行列号相同的特征点分为一组；按照海冰的最大漂移速度计算匹配半径，对两幅影像特征点进行匹配；The third module is used to match the feature points of the two images by using the matching method based on geographic grid; the matching method based on geographic grid is realized by grouping the feature points on the images according to the geographic grid , calculate the row and column number of the grid where each feature point is located, and group the feature points with the same row and column number into a group; calculate the matching radius according to the maximum drift speed of the sea ice, and match the two image feature points;
第四模块，用于根据匹配点对的位置以及两幅影像的时间差，计算初始海冰运动矢量；The fourth module is used to calculate the initial sea ice motion vector according to the position of the matched point pair and the time difference between the two images;
第五模块，用于根据初始海冰运动矢量，按照指定空间分辨率规则地内插海冰运动矢量；The fifth module is used to regularly interpolate the sea ice motion vector according to the specified spatial resolution according to the initial sea ice motion vector;
第六模块，用于针对内插所得海冰运动矢量和真实的海冰运动矢量存在的偏差，按照最大互相关方法进行精化；The sixth module is used to refine the deviation between the sea ice motion vector obtained by interpolation and the real sea ice motion vector according to the maximum cross-correlation method;
第七模块，用于采用局部一致性滤波对精化后的海冰运动矢量场进行滤波，剔除错误的海冰运动矢量，获得海冰运动提取结果。The seventh module is used to filter the refined sea ice motion vector field by using local consistency filtering, remove the wrong sea ice motion vector, and obtain the sea ice motion extraction result.
在一些可能的实施例中，提供一种适用于SAR和光学影像的高效海冰运动提取系统，包括处理器和存储器，存储器用于存储程序指令，处理器用于调用存储器中的存储指令执行如上所述的一种适用于SAR和光学影像的高效海冰运动提取方法。In some possible embodiments, an efficient sea ice motion extraction system suitable for SAR and optical imaging is provided, comprising a processor and a memory, the memory is used to store program instructions, and the processor is used to call the stored instructions in the memory to execute the above An efficient sea ice motion extraction method suitable for SAR and optical images.
在一些可能的实施例中，提供一种适用于SAR和光学影像的高效海冰运动提取系统，包括可读存储介质，所述可读存储介质上存储有计算机程序，所述计算机程序执行时，实现如上所述的一种适用于SAR和光学影像的高效海冰运动提取方法。In some possible embodiments, an efficient sea ice motion extraction system suitable for SAR and optical imaging is provided, including a readable storage medium, on which a computer program is stored, and when the computer program is executed, An efficient sea ice motion extraction method suitable for SAR and optical images as described above is realized.
本文中所描述的具体实施例仅仅是对本发明精神作举例说明。本发明所属技术领域的技术人员可以对所描述的具体实施例做各种各样的修改或补充或采用类似的方式替代，但并不会偏离本发明的精神或者超越所附权利要求书所定义的范围。The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which the present invention pertains can make various modifications or additions to the described specific embodiments or substitute in similar manners, but will not deviate from the spirit of the present invention or go beyond the definitions of the appended claims range.
Claims (8)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210178701.9A CN114612773B (en) | 2022-02-25 | 2022-02-25 | Efficient sea ice motion extraction method and system suitable for SAR and optical images |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210178701.9A CN114612773B (en) | 2022-02-25 | 2022-02-25 | Efficient sea ice motion extraction method and system suitable for SAR and optical images |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114612773A true CN114612773A (en) | 2022-06-10 |
| CN114612773B CN114612773B (en) | 2024-06-25 |
Family
ID=81858318
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210178701.9A Active CN114612773B (en) | 2022-02-25 | 2022-02-25 | Efficient sea ice motion extraction method and system suitable for SAR and optical images |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114612773B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114937145A (en) * | 2022-07-26 | 2022-08-23 | 北京数慧时空信息技术有限公司 | Remote sensing image feature point matching method based on geological information |
| CN115524343A (en) * | 2022-09-29 | 2022-12-27 | 哈尔滨工业大学 | Microscopic characterization method of physical structure of ice crystal |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB8823588D0 (en) * | 1988-10-07 | 1988-11-16 | Marconi Gec Ltd | Tracking sea ice |
| WO2006075394A1 (en) * | 2005-01-14 | 2006-07-20 | Morhpo, Inc. | Motion vector calculation method, hand-movement correction device using the method, imaging device, and motion picture generation device |
| CN110956083A (en) * | 2019-10-21 | 2020-04-03 | 山东科技大学 | Bohai sea ice drift remote sensing detection method based on high-resolution four-signal optical satellite |
| CN112906645A (en) * | 2021-03-15 | 2021-06-04 | 山东科技大学 | Sea ice target extraction method with SAR data and multispectral data fused |
-
2022
- 2022-02-25 CN CN202210178701.9A patent/CN114612773B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB8823588D0 (en) * | 1988-10-07 | 1988-11-16 | Marconi Gec Ltd | Tracking sea ice |
| WO2006075394A1 (en) * | 2005-01-14 | 2006-07-20 | Morhpo, Inc. | Motion vector calculation method, hand-movement correction device using the method, imaging device, and motion picture generation device |
| CN110956083A (en) * | 2019-10-21 | 2020-04-03 | 山东科技大学 | Bohai sea ice drift remote sensing detection method based on high-resolution four-signal optical satellite |
| CN112906645A (en) * | 2021-03-15 | 2021-06-04 | 山东科技大学 | Sea ice target extraction method with SAR data and multispectral data fused |
Non-Patent Citations (1)
| Title |
|---|
| 万雷等: "基于InSAR和ICESat的南极冰盖地区DEM提取和精度分析", 《冰川冻土》, vol. 37, no. 05, 31 December 2015 (2015-12-31) * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114937145A (en) * | 2022-07-26 | 2022-08-23 | 北京数慧时空信息技术有限公司 | Remote sensing image feature point matching method based on geological information |
| CN114937145B (en) * | 2022-07-26 | 2022-09-20 | 北京数慧时空信息技术有限公司 | Remote sensing image feature point matching method based on geological information |
| CN115524343A (en) * | 2022-09-29 | 2022-12-27 | 哈尔滨工业大学 | Microscopic characterization method of physical structure of ice crystal |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114612773B (en) | 2024-06-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112232134B (en) | Human body posture estimation method based on hourglass network and attention mechanism | |
| CN114612773B (en) | Efficient sea ice motion extraction method and system suitable for SAR and optical images | |
| CN113963117B (en) | Multi-view three-dimensional reconstruction method and device based on variable convolution depth network | |
| CN113112547B (en) | Robot, repositioning method thereof, positioning device and storage medium | |
| CN103871039B (en) | Generation method for difference chart in SAR (Synthetic Aperture Radar) image change detection | |
| CN109741446B (en) | Method for dynamically generating fine coast terrain by three-dimensional digital earth | |
| CN107103610B (en) | automatic detection method for suspicious region matched with satellite images in stereo mapping | |
| CN116703996A (en) | Monocular 3D Object Detection Algorithm Based on Instance-Level Adaptive Depth Estimation | |
| CN112785629A (en) | Aurora motion characterization method based on unsupervised deep optical flow network | |
| CN113112522A (en) | Twin network target tracking method based on deformable convolution and template updating | |
| CN112132880A (en) | A real-time dense depth estimation method based on sparse measurements and monocular RGB images | |
| CN109544609A (en) | A kind of sidescan-sonar image matching process based on SIFT algorithm | |
| CN114581864A (en) | Transformer-based dynamic densely aligned vehicle re-identification technology | |
| CN111696167A (en) | Single image super-resolution reconstruction method guided by self-example learning | |
| CN115082778B (en) | Multi-branch learning-based homestead identification method and system | |
| Tang et al. | Feature extraction for side scan sonar image based on deep learning | |
| CN116193103A (en) | Video picture jitter level assessment method | |
| CN103559723B (en) | A kind of human body tracing method based on self-adaptive kernel function and mean shift | |
| CN118608574B (en) | A visual odometer method, storage medium and device based on illumination suppression | |
| CN115826089B (en) | East-west vortex boundary extraction method based on vortex tracking data | |
| CN115346004B (en) | Remote sensing time sequence data reconstruction method combining space-time reconstruction and CUDA acceleration | |
| Zou et al. | Multi-dimensional attention unet with variable size convolution group for road segmentation in remote sensing imagery | |
| Cai et al. | LiDAR point cloud image interpolation via separable convolution | |
| CN114743013B (en) | Local descriptor generation method, device, electronic device and computer program product | |
| CN119206149B (en) | A 3D point cloud data completion method based on multimodal joint perception |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |