205 results about "Remote sensing image segmentation" patented technology

A remote sensing image segmentation method combining complete residual and multi-scale feature fusion includes S100 improving a convolution coding-decoding network as a segmentation backbone network,separately comprising S101 using the convolution coding-decoding network as the segmentation backbone network; S102 adding a feature pyramid module for aggregating multi-scale context information intothe backbone network; S103 adding a residual unit into the convolution layer corresponding to the encoder and the decoder of the backbone network, and meanwhile, fusing the features in the encoder into the corresponding layer of the decoder in a pixel-by-pixel manner; S200 using the improved image segmentation network combined with complete residual and multi-scale feature fusion for remote sensing image segmentation; S300 outputting the segmentation result of the remote sensing image. This method not only simplifies the training of the deep network and enhances the feature fusion, but also enables the network to extract rich context information, cope with changes in the scale of the target, and improve the segmentation performance.

The invention discloses a semi-supervised multi-spectral remote sensing image segmentation method based on spectral clustering; the segmentation process includes that: (1) the characteristics inputted to the multi-spectral sensing image are extracted; (2) N points without labels and M points with labels are randomly and evenly sampled from a multi-spectral sensing image with S pixel points to form a set n which is the summation of N and M, wherein M points with labels are used for creating pairing limit information Must-link and Cannot-link sets; (3) the sampled point set is analyzed through semi-supervised spectral clustering to obtain the class labels of the n (n=N+M) points; (4) the sampled n (n=N+M) points are used as the training sample to classify the rest (S-N-M) points through nearest-neighbor rule, each pixel point is assigned with a class label according to the class of the pixel point and is used as the segmentation result of the inputted image. Compared with prior art, the invention has good image segmentation effect, strong operability, improves the classification accuracy, avoids searching the optimum parameters through repeated test, has small limit on image size and is better applicable to the segmentation of multi-class multi-spectral sensing images.

The invention discloses a remote sensing image segmentation method based on region clustering, belonging to the field of remote sensing image comprehensive utilization. The method comprises the following steps: carrying out region pre-segmentation by a MeanShift algorithm to remove noise and perform initial cluster on image elements; carrying out fuzzy clustering on images which are subject to the pre-segmentation by a fuzzy C-means algorithm (FCM), and initially inducing and identifying characteristics of each image object to obtain the probability that each object affiliates to some a category so as to constitute a land category probability space of the remote sensing images, thereby providing a basis of object combination for further region segmentation; and performing region segmentation in the probability space of clustering images, classifying image elements which are close in the probability space and similar in the category as the same objects by region labels. In the method of the invention, two defects in the existing segmentation method are overcome, the remote sensing images can be effectively and accurately segmented, segmentation tasks of the remote sensing images can be finished by batch by integration, and data support can be preferably provided for extraction of land information from the remote sensing images.

The invention discloses a remote sensing image ship integrated recognition method based on deep learning. The method comprises the steps of image classification, target detection and image segmentation. Compared with the prior art, the method has the advantages that the modern artificial intelligence deep learning model is combined with the traditional image processing method to detect and segmentthe ship with the eye remote sensing image; The remote sensing image segmentation method based on deep learning can accurately identify ships in a sea area, is suitable for various processing environments, has a good anti-interference capability for a complex environment, and can accurately segment the detected and identified ships.

The invention relates to a method for segmenting a remote sensing image and belongs to the technical field of processing of the remote sensing image. The method comprises the following steps of: according to color feature, textural feature and the position information of a pixel point, generating a feature vector which serves as an initial input vector; segmenting the remote sensing image by using mean shift filtering; and adaptively determining weights of the color feature and the textural feature by using constructed cluster validity indexes to effectively improve the segmenting accuracy ofthe remote sensing image. The invention also discloses a method for searching the remote sensing image by using the method for segmenting the remote sensing image. The search of the remote sensing image is finished by using a new region representation and matching method. Compared with the prior art, the invention has higher search accuracy, particularly for search of a small target region.

The invention provides a cloud accessing method for mass remote sensing data. A cloud computing platform of the remote sensing data is established based on Hadoop, a remote sensing image pyramid of the remote sensing data is then established according to the resolution ratio of the remote sensing data, each remote sensing image is divided into a plurality of slices according to the remote sensing image pyramid, all the slices are combined to form integrate images, the images are written into a newly established HDFS, a remote sensing data cloud storage module is further established by adopting the mode of a pyramid image storage structure, and a remote sensing data cloud reading module is established. According to the cloud accessing method for the mass remote sensing data, the cloud computing is introduced into the method, the slicing and combination of the mass remote sensing images are carried out based on the cloud computing platform, the mass remote sensing data cloud storage module and the mass remote sensing data cloud reading module are established, and fast and effective storage and reading can be carried out on the mass remote sensing data.

The invention discloses a multi-band high-resolution remote sensing image segmentation method based on a gray scale co-occurrence matrix. In texture images based on the gray scale co-occurrence matrix, precipitation watershed transformation is adopted for segmentation of each band image, and therefore band segmentation results are superposed. At last, a region merging strategy based on multi-band spectral information is provided for merging fragment regions in the segmentation results, and finally image segmentation is achieve. A high-resolution ALOS image and an SPOT 5 image are tested respectively through the method which is compared with a traditional segmentation method based on local area homogeneity gradients. As is shown in experimental results, edges of an object can be accurately positioned, over-segmentation and under-segmentation are effectively overcome, and higher segmentation accuracy and stability are achieved.

The invention provides a remote sensing image segmentation and identification method based on superpixel marking. Superpixel segmentation results are obtained by performing over-segmentation on remote sensing images by use of a superpixel segmentation algorithm, and learning samples are obtained by performing classification marking on superpixel blocks. Then, visual features of superpixel samples are extracted, these learning samples are trained by taking marking results as teacher signal classifiers, and trained classifier information is stored. The superpixel results are obtained by performing the over-segmentation on the remote sensing images to be analyzed, a visual feature of each superpixel is extracted and then is sent to the classifiers for classification, and after each superpixel block obtains a class mark, the superpixel blocks of the same class marks are merged, i.e., all areas of the images to be analyzed obtain class information. According to the invention, the remote sensing images are prevented from being directly segmented, edge information of actual areas is reserved to a quite large degree, segmentation and identification processes are integrated together, and the segmentation and identification capabilities are more excellent.

The invention relates to a method and a device for processing a remote sensing image. The method comprises the following steps: obtaining the largest data processing amount of a plurality of subnodes in the grid; segmenting the remote sensing imaged into a plurality of remote sensing subimages according to the largest data processing amount; and sending the remote sensing subimages to the subnodes and then leading the subnodes to respectively process the received remote sensing subimages. The method and the device for processing the remote sensing image lead the remote sensing image to be segmented into the remote sensing subimages according to the largest data processing amount of the subnodes in the grid, therefore, the subnodes in the grid can parallelly process the remote sensing subimages according to the processing capacity of the subnodes, thus increasing the processing speed of the remote sensing image.

The invention discloses a high-resolution remote sensing image segmentation method based on inter-scale mapping. The high-resolution remote sensing image segmentation method has the advantages that aiming at geographical object extraction involved in objet-level change detection, the high-resolution remote sensing image multi-scale segmentation method based on wavelet transform and an improved JSEG (joint systems engineering group) algorithm is provided; aiming at a key problem that a conventional JSEG algorithm affects segmentation accuracy in high-resolution remote sensing image segmentation, a corresponding improvement strategy is adopted to achieve a good effect; wavelet transform is introduced to serve as a multi-scale analysis tool, excessive rough color quantization of the conventional JSEG algorithm is abandoned, and thus, detail information in original images is kept to the utmost; a novel inter-scale segmentation result mapping mechanism is set up, image segmentation of a current scale is realized on the basis of a segmentation result of a previous scale, and correction of the segmentation result of the previous scale is also realized, and thus, accumulation of inter-scale segmentation errors is decreased effectively; finally, an improved multi-scale segmentation strategy and an improved region merging strategy are provided, and over-segmentation and mistaken merging are effectively reduced.

The invention discloses a method and a system for change detection based on a multi-temporal remote sensing image and GIS data. The method is aimed at solving problems existing in segmentation, classification and change detection methods based on multi-temporal images and GIS data. The method includes the following steps: step 1, segmentation of a remote sensing image and extraction of pixel-levelchange information, mainly including remote sensing image segmentation and pixel-level change detection; step 2, object-oriented remote sensing image classification, mainly including change polygon decision, automatic sample selection and object-oriented SVM decision tree classification model; step 3, change detection after the object-oriented classification. Compared with existing methods of extracting change information using remote sensing images and GIS data, the method can realize automatic and large-area application change information detection method flow, can fully excavate relationship between remote sensing images and GIS data, utilizes attribute information such as type of the GIS data, and improves accuracy of change information extraction.

The invention discloses a multi-scale remote-sensing image segmentation method based on the local homogeneity index. The segmentation method comprises image quantization based on the bisecting K-means clustering, multi-scale segmentation based on surface feature context information and region merging based on SSIM and D-S evidence theory. According to the invention, coarse image quantization in JSEG is improved; the multi-scale segmentation based on the surface feature context information and the SSIM and the region merging strategy based on the D-S evidence theory are provided; by carrying out segmentation experiments on multiple groups of different sensor types of high-definition remote sensing images, it can be proved that the boundary of objectives can be precisely positioned via the disclosed algorithm; and the segmentation method is provided with quite high segmentation precision.

The invention relates to a high-resolution remote sensing image segmentation method based on texture clustering constraint in the field of remote sensing technology, which comprises the concrete steps of: step 1, calculating Gabor energy textures of all regions in an image, carrying out clustering of all the regions by using FCM according to texture distance, and setting a texture clustering tag for each region according to clustering results; step 2, setting up a comprehensive distance space model by using features such as spectrum, texture, shape and the like, and adding texture clustering distance to restrict and lead region merging to be performed along the homogeneous texture direction; and step 3, setting up an RAG model and an NNG model according to the comprehensive distance, and merging the regions according to global optimum. The real regional boundary is obtained by the interaction of the texture clustering and optimal merging sequence during the merging. The method can better segment the texture region in the high-resolution image and improve the overall segmentation accuracy of the image.

The invention relates to a watershed algorithm-based high spatial resolution multi-spectral remote sensing image segmentation method, and aims to solve the segmentation problem of a high spatial resolution multi-spectral remote sensing image. The method comprises the following steps of: solving the gradient of the multi-spectral image by using a multi-spectral gradient algorithm to obtain a gradient image, then segmenting the gradient image by using a watershed algorithm, and finally performing region merging according to a merging algorithm by using the similarity between regions to obtain a final segmentation result.

A method for rebuilding super resolution of small object on single frame satellite remote sense image includes setting up clear - drop model of on - track satellite remote sense influence, discriminating clear - drop parameter, estimating S - N ratio and calculating normalization factor, cutting said image to be superposed sub blocks by rebuilding algorithm, normalizing frequency domain and redundant wavelet domain, decorrelating de - noising operator, setting up fast algorithm of super resolution image rebuild, rebuilding sub pixel, using modulation transfer function as objective standard for evaluating resolution and quality of image.

The invention discloses a quantum multi-target clustering-based remote sensing image segmentation method, which mainly solves the problems of single evaluation index, high calculation complexity and poor segmentation effect in the conventional image segmentation technology. The quantum multi-target clustering-based remote sensing image segmentation method comprises the following implementation steps of: (1) inputting a remote sensing image to be segmented; (2) extracting the characteristics of the image to be segmented; (3) generating clustering data; (4) randomly generating an initial quantum population to finish initialization; (5) acquiring a binary population; (6) calculating an individual fitness value; (7) selecting non-dominated sorting; (8) evolving the population; (9) judging whether to meet stop condition; (10) distributing a category label; (11) generating the best individual; and (12) outputting a segmentation image. According to the quantum multi-target clustering-based remote sensing image segmentation method, the characteristics of each pixel of the image are extracted; remote sensing image segmentation is realized through a quantum calculation and multi-target optimization combined clustering method; and the quantum multi-target clustering-based remote sensing image segmentation method has the advantages of high segmentation accuracy and accurate edge positioning and can be used for segmenting a complex image.

The invention discloses a high resolution remote sensing image segmentation method based on Gram-Schmidt fusion and locally excitatory globally inhibitory oscillator networks (LEGION). The method comprises the following steps: resampling the multispectral wave bands of a high resolution remote sensing image, and enabling the multispectral wave bands to have same size and pixel amount as a panchromatic wave band; performing Gram-Schmidt fusion on the panchromatic wave band and the multispectral wave bands to lead each of the multispectral wave bands to have higher space resolution and basically maintain the spectral information; calculating the average value of the all wave band pixel values of all pixels, and merging the information of the multiple wave bands into a wave band to serve as input data of an LEGION segmentation method; segmenting the merged single wave band image by the LEGION method; and inputting the segmentation result in a result image, and displaying the result image in a visualized mode. By means of the method in the invention, the defect that the LEGION segmentation method only can utilize single wave band information is solved, and the high resolution remote sensing image can be segmented more accurately and effectively.

The invention relates to a multi-scale feature fusion remote sensing image segmentation method and device, equipment and a memory. The method comprises the following steps: collecting a remote sensing image, and marking to obtain a training sample; a multi-scale feature fusion remote sensing image segmentation network is constructed, and the network comprises an input network which is used for segmenting a training sample into small blocks with fixed sizes, expanding the small blocks into one-dimensional vectors and embedding position codes to obtain an input sequence; the encoder is used for extracting different levels of an input sequence by utilizing a multi-layer Transform module; the decoder is used for prediction result by fusing the multi-scale feature map; and training the network by using a training sample to obtain a trained multi-scale feature fusion remote sensing image segmentation model, and obtaining a prediction result of a to-be-detected remote sensing image by using the model. According to the method, a multi-scale feature map extracted by an encoder is fully utilized, local classification and hierarchical segmentation are combined, and the method can adapt to the characteristic that targets in remote sensing images are complex and changeable.

The invention discloses a remote sensing image segmentation method based on a feature recalibration dual-channel residual network, and the method comprises the steps: S1, obtaining a to-be-detected remote sensing image, and carrying out the normalization processing of the to-be-detected remote sensing image, so as to obtain a normalized image; S2, performing random window sampling on the normalized image to obtain a plurality of sample data with set sizes; S3, segmenting the plurality of sample data by adopting a pre-established dual-channel residual network remote sensing segmentation model to obtain segmented remote sensing images. The training process of the dual-channel residual network remote sensing segmentation model is that the dual-channel residual network is trained according toa known remote sensing image to obtain the dual-channel residual network remote sensing segmentation model.

The invention relates to the field of image processing, and discloses a quick high-resolution remote sensing image segmentation method. The method comprises steps: S1, the high-resolution remote sensing image is read; S2, a multiband morphological gradient for the read high-resolution remote sensing image is calculated to obtain a multiband morphological gradient image formed by the multiband morphological gradient; S3, morphological rebuilding is carried out on the multiband morphological gradient image to obtain a gradient image after morphological rebuilding; S4, watershed segmentation is carried out on the gradient image after morphological rebuilding to obtain an image after watershed segmentation; and S5, region merging is carried out on the image after watershed segmentation. The remote sensing image with mass data can be quickly segmented, and over segmentation can be effectively reduced.

The invention relates to the technical field of image detection and image classification, and provides a remote sensing image segmentation method, which comprises the following steps: obtaining a target image, and extracting hierarchical characteristics of the target image through each convolutional layer of a deep convolutional neural network; determining target features of the target image basedon the hierarchical features extracted from the convolution layers of the two different scales in sequence; and determining the object category of each object in the target image according to the target feature, and segmenting the target image according to the object category. According to the invention, the convolutional neural network is adopted to extract multi-scale features from shallow to deep; and deep, middle and shallow features are fused in a jumping mode to obtain high-level semantic information and low-level geometric information, and then spatial output is realized through up-sampling, so that the purpose of pixel level segmentation is achieved, and the segmentation precision of the remote sensing image under cloud, strong illumination and other interference factors is improved.

The invention discloses an unmanned aerial vehicle remote sensing image segmentation method based on color space classification. According to the method, a high-resolution unmanned aerial vehicle remote sensing image is divided into a series of pixel clusters. After that, pixel points are classified based on the similarity of color space coordinates and texture features. Finally, the fusion of pixel clusters is guided in combination with the classification result. As a result, it is guaranteed that each partition basically belongs to a homogeneous region. The problems of low classification precision, large pepper salt noise and low algorithm efficiency existing in the prior art are solved. The purpose of quickly and accurately segmenting the high-resolution unmanned aerial vehicle remote sensing image is realized.

The present invention discloses an urban high-resolution remote sensing image shadow detection and segmentation method. The method includes the following steps that: image quantification is performedon urban multi-band remote sensing images, shadow detection and compensation are performed on the quantized remote sensing images on the basis of chi-square transformation, so that shadow compensatedimages can be obtained; and a multi-scale J-image sequence is calculated for the obtained shadow compensated images, and multi-scale segmentation and region combination are performed, so that final remote sensing image segmentation results can be obtained. With the method of the invention adopted, weak edges and false edges caused by shadows can be can effectively dealt with. The segmentation accuracy of the method is significantly improved; and the method has high reliability.

The invention relates to an agricultural remote-sensing image segmentation technology, in particular to a near-rectangle guide based remote-sensing cornfield image segmentation method. The method includes: subjecting a fused satellite remote-sensing image to edge extraction by an SUSAN (small univalue segment assimilating nucleus) operator, then constructing a near-rectangle guide based correlation function according to a relation between a closed region and a circumscribed near-rectangle, and importing a near-rectangle threshold function into an image-based segmentation algorithm to segment a field in a specific shape. According to the method, a satisfactory segmentation result can be obtained to better conform to an actual cornfield area statistic result, and over-segmentation patches of boundary regions due to a same object with different spectra are decreased. The near-rectangle guide based remote-sensing cornfield image segmentation method is mainly used for remote-sensing measurement of large-area corn planting area.

The present invention discloses a target level remote sensing image change detection method based on a RBM model, mainly aims at the defects of an existing change detection method, combines a Restricted Boltzmann Machine (RBM) with target level remote sensing image change detection and is applied to remote sensing image change detection. The target level remote sensing image change detection method comprises the implementing steps of: (1) inputting gray matrices of two remote sensing images; (2) carrying out fuzzy clustering to obtain the segmented gray matrices of two remote sensing images; (3) constructing a logarithm ratio difference gray matrix to be detected; (4) carrying out pre-classification on the logarithm ratio difference gray matrix; (5) selecting a training sample; (6) training the RBM; and (7) outputting a change detection result. According to the present invention, the dependency on remote sensing image registration accuracy is reduced; the target level remote sensing image change detection method has the excellent noise immunity; and accuracy and classification accuracy of remote sensing image change detection are improved.

The invention relates to an image processing technology, in particular to an automatic optimization method of initially segmented boundary based on neighbor function criterion. The method comprises the following steps: initially segmented boundary points are detected, and are marked with Arabic numerals according to an initially segmented area; neighbor function values within a neighborhood are calculated for each boundary point; membership function values of the current boundary point are calculated according to the neighbor function values; the boundary points are reclassified according to the membership function values to get an optimized segmented boundary; the steps above are repeated to ensure the segmented boundaries of the entire image to be optimized. The method provided by the invention imitates some functions of human eyes during image processing, and can optimize inaccurately segmented boundaries automatically. In addition, the invention eliminates the influence of image noise, local bulk effect, overlapping intensity and non-uniformity of intensity, and well complements prior segmentation algorithm. The invention has important application values in the fields of medical image segmentation, remote sensing image segmentation, target identification and so on.

The invention discloses a remote sensing image segmentation method based on a disparity map and a multi-scale deep network model, which mainly solves the problems of low segmentation precision and weak robustness of the existing remote sensing image segmentation method. The implementation scheme includes: reading a data set, and generating a training data set of remote sensing image segmentation;constructing a multi-scale fusion segmentation network model; using the training data set to train a segmentation network model, and storing seven models with different iteration times; obtaining seven different segmentation result graphs by using the stored segmentation network model; carrying out majority voting on the seven different segmentation result graphs, and carrying out super-pixel processing on the voted result graph to obtain a preliminary segmentation result graph; obtaining a disparity map of the test scene by using an SGBM algorithm; and optimizing the preliminary segmentationresult graph by using the disparity map to obtain a final segmentation result. Compared with an existing method, the method has the advantages that the segmentation precision and robustness are obviously improved, and the method can be widely applied to urban and rural planning and intelligent urban construction.

The invention discloses a river and lake remote sensing image segmentation method and system based on a convolutional neural network and Transform, and the method comprises the steps: obtaining a river and lake remote sensing image containing a class label, and constructing a training set; extracting a multi-layer feature map from the training set by using a convolutional neural network; extracting an attention feature from the last layer of extracted feature map by adopting a Transform encoder, and obtaining a self-attention feature map from the attention feature by adopting a Transform decoder; after the self-attention feature map and the first-layer feature map are spliced, training an image segmentation model; and obtaining a category segmentation result of the target in the remote sensing image of the river and lake to be detected based on the trained image segmentation model. Transform is introduced into the field of remote sensing image segmentation, a self-attention mechanism is used for replacing convolution operation, the receptive field area during operation is enlarged, due to the fact that down-sampling and up-sampling operation does not exist, image scale change cannot be caused, the problem of target loss is solved, and the defects of an existing deep learning segmentation method in the field of remote sensing image segmentation are overcome.

The invention provides a high-resolution remote sensing image segmentation method. The method comprises steps that a to-be-segmented high-resolution remote sensing image is read; a Gaussian type-2 fuzzy membership function model of each ground feature category is utilized to calculate a Gaussian type-2 fuzzy membership degree corresponding to each gray level; a segmentation decision model of each ground feature category is utilized to calculate a membership degree of each gray level in each segmentation decision model; a ground feature category of a largest membership degree value of a gray level of each pixel in the segmentation decision models is a segmentation result; the Gaussian type-2 fuzzy membership function model is changed according to set step length, all the segmentation results are compared, a segmentation result with highest segmentation precision is taken as a final high-resolution remote sensing image segmentation result. According to the method, a segmentation problem caused by gray level membership uncertainty and segmentation decision uncertainty can be effectively solved, more precise fitting of high-resolution remote sensing data complex histogram distribution characteristics is realized, a noise problem is eliminated, and segmentation precision is improved.

The invention discloses a remote sensing image partition method based on an automatic difference clustering algorithm. The method mainly solves the problems in the existing image partition technology of being high in calculating complexity and poor in partition effect. The remote sensing image partition method includes the steps: (1) inputting an image to be partitioned and extracting features of the image to be partitioned; (2) generating clustering data; (3) drawing clustering data initial population randomly; (4) activating a clustering center according to individual labels; (5) calculating an individual fitness value according to the activated clustering center; (6) evolving the population through an improved difference evolving method; (7) conducting oscillation operation of the number of categories on the evolved population; (8) updating a center of mass by using a fuzzy C means (FCM); (9) judging end conditions by using the updated center of mass and recording the optimal individuals; and (10) decoding the optimal individuals, distributing category labels and outputting partitioned images. The method has the advantages of being high in partition precision and accurate in border locating and can be used for target identification.