41 results about "Change detection algorithms" patented technology

The invention provides an SAR (synthetic aperture radar) image change detection algorithm based on non-supervision depth network learning. The algorithm includes the steps: 101 starting an SAR image change detection method based on a non-supervision depth nerve network; 102 performing FCM (fuzzy c-mean) joint classification for two registered SAR images of different time phases in the same area to obtain rough change detection results; 103 selecting noiseless points with large possibility to serve as training samples of the depth network according to initial change detection results; 104 inputting sample points to be trained to the designed depth nerve network to be trained; 105 inputting two images to be detected to the trained depth nerve network to obtain a final change detection result map; 106 finishing the SAR image change detection method based on the non-supervision depth nerve network. Ohm= {ohm1 and ohm2}. Construction links of a difference map are avoided, sensitivity of noise is improved to a certain extent, and detection efficiency and detection accuracy are remarkably improved.

The invention discloses a change detection method based on a high-resolution remote sensing image. The method is characterized by using a superpixel segmentation and synthesis algorithm to segment a multi-temporal remote sensing image after carrying out necessary ortho-rectification, image registration, histogram matching and other preprocessing, using a superpixel as a unit to carry out local characteristic calculating and sample selection, and realizing the automatic annotation of a changing area possessing an obvious tendency or an unchanged area in the image; and then, taking an annotationresult as a sample to train twinborn convolutional neural networks, classifying image changing conditions, carrying out noise reduction, morphological filtering and other postprocessing, and acquiring a final changing detection result. An experiment shows that in a high-score number-two satellite remote sensing image data set, each index in the method is better than each index in a traditional change detection algorithm; a Kappa coefficient is averagely increased by 0.3; and an average overall error rate is less than 3.5%. The detection result has a high practical value.

The invention discloses a disposal method used for the digital delay of an active power filter and the corresponding devices. The invention comprises the following steps that a load current is sampled by selecting a time point which is prior to the time Vt when duty cycle loading is implemented, and the current value of the load at current time is detected; the harmonic component in the load current is calculated based on a change detection algorithm; the output current of active power filter is sampled by selecting the middle-point times of all switch periods of PWM signals; close ring control is carried out for the harmonic component of the load current and the output current of the active power filter, the PWM signal is modulated and a power device is driven. By adopting the method, the effect of digital delay to the compensation precision of the active power filter is reduced; the sampling delay of the load current with periodic harmonic is eliminated; the operation time is little, and full digitization is easy to be implemented. The invention is more applicable to the microprocessor chips with medium prices and processing speed, and reduces the cost of active power filter.

The invention discloses a quick landslide extraction method based on fully polarimetric SAR (synthetic aperture radar) images. An airborne or satellite-borne satellite acquires two to-be-processed original fully polarimetric images with the same orbit and mode before and after occurrence of a landslide disaster, then reference images are acquired through preprocessing of the images, polarization feature extraction is performed on the reference images with a Freeman decomposition and Cloude decomposition combined method, initial classification images are acquired and then clustered, similar feature areas are merged, final classification images are acquired, change detection processing is performed on the final classification images with a change detection algorithm according to area features before and after landslide occurrence, a suspected landslide area chart is generated, finally, the suspected landslide area chart is processed with a morphological method, noise or small areas are removed, and a landslide distribution map is output finally, so that large-area landslide information can be extracted from the fully polarimetric SAR images quickly and accurately.

According to the unmanned aerial vehicle aerial image change detection algorithm based on semantic segmentation, an unmanned aerial vehicle is used for shooting the same place in two different time periods to collect image data, an original image is obtained, and the algorithm is characterized by comprising the following steps that S1, making a label coding set; S2, making a training set; S3, training to generate a semantic segmentation network; S4, obtaining two semantic segmentation result images; and S5, obtaining a change detection result. According to the algorithm, the semantic level characteristics of the image are fully utilized, some auxiliary training sets are constructed for training, so that the trained network can learn some generalized characteristics, and the final detectioneffect is higher than that of the traditional manual characteristic. According to the method, the final change result can judge whether the change occurs or not, the change category can be detected,16 change categories can be recognized at present, and some requirements for change detection in actual research problems are better met.

The invention provides an on-line linkage update service method of multi-scale geographic data in an internet mode. A multi-scale geographic data linkage back-trace expansion updating mechanism is achieved, and geographic ontology is applied to establish a multi-scale geographic data presentation model and construct a multi-scale geographic special database. A characteristic variable quantity of different-phase multi-source data is discovered by applying change detection algorithm, an incremental change data package and ground state geographic data are subjected to element-level on-line update to form uptodate state geographic data, and the incremental change data package is expanded and updated to other-scale geographic data by multi-scale data integration and service composition of spatial information services. By the on-line linkage update service method of the multi-scale geographic data, integration and sharing of the multi-source heterogeneous multi-scale geographic database are guaranteed while on-line update, linkage update and dynamic update of service-oriented distributed geographic information processing are achieved.

A system and method for determining the exact pair of alleles corresponding to polymorphic genes from sequencing data and for using the polymorphic gene information in formulating an immunogenic composition. Reads from a sequencing data set mapping to the target polymorphic genes in a canonical reference genome sequence, and reads mapping within a defined threshold of the target gene sequence locations are extracted from the sequencing data set. Additionally, all reads from the set data set are matched against a probe reference set, and those reads that match with a high degree of similarity are extracted. Either one, or a union of both these sets of extracted reads are included in a final extracted set for further analysis. Ethnicity of the individual may be inferred based on the available sequencing data which may then serve as a basis for assigning prior probabilities to the allele variants. The extracted reads are aligned to a gene reference set of all known allele variants. The allele variant that maximizes a first posterior probability or posterior probability derived score is selected as the first allele variant. A second posterior probability or posterior probability derived score is calculated for reads that map to one or more other allele variants and the first allele variant using a weighting factor. The allele that maximizes the second posterior probability or posterior probability score is selected as the second allele variant.A system and method for identifying somatic changes in polymorphic loci using WES data. The exact pair of alleles corresponding to the polymorphic gene are determined as described using a normal or germline sample from an individual. A tumor or otherwise diseased sample is also retrieved from the individual and the corresponding WES data is generated. Reads corresponding to the polymorphic gene are extracted as described in the paragraph above. These reads are then aligned to the inferred pair of allele sequences. The alignment of the germline or normal reads to the inferred pair of alleles, along with the alignment of the tumor or diseased reads to the inferred pair of alleles are simultaneously used as inputs to somatic change detection algorithms to identify somatic changes with greater precision and sensitivity.

The invention discloses a CPU+FPGA-based rapid SAR image change detection method. According to the method, change detection of SAR images is realized by adoption of CPU+FPGA heterogeneous parallel acceleration; and compared with SAR image change detection methods in the prior art, the method has the advantage that the parallel calculation ability of the FPGA is much higher than the serial data calculation ability of the CPU. An FPGA and CPU combined work manner is adopted, the FPGA is responsible for enabling the data processing in the whole algorithm to be parallel, and the CPU is responsiblefor controlling task scheduling and parameter size at a host side. According to the method, parallel programming is carried out on an SAR image change detection algorithm by adoption of a parallel programming standard of OpenCL, and codes are relatively high in portability and can be operated in different heterogeneous platforms.

The invention discloses a synthetic aperture radar (SAR) image change detection method based on a multi-objective evolutionary algorithm based on decomposition (MOEA / D) and fuzzy clustering. The problem of balancing two objectives, namely detail maintenance and noise removal in SAR image change detection is solved by a multi-objective optimization method. The method comprises the following steps of (1) generating a difference chart by using a logarithm contrast method for an image to be detected; (2) filtering the difference chart to obtain the denoised difference chart; (3) determining two objective functions according to the two objectives of detail maintenance and noise removal, and combining into a multi-objective optimization problem; (4) obtaining a Pareto front end of the multi-objective problem and a corresponding result chart by using the MOEA / D; (5) selecting a suitable change detection result chart from all results according to a requirement. Compared with a condition that only one solution is obtained by other change detection algorithms, the method has the greatest advantage that the method is used for obtaining an optimal solution set, and a user can select a more suitable solution according to the emphasis on the detail maintenance and the noise removal.

The invention provides a remote sensing image data space-time fusion method, system and device. The method includes obtaining a change detection imagethrough calculation of two time-phase low-resolution remote sensing images; extracting an edge region of the high-resolution image of the first time phase, and calculating abundance corresponding to the number of various high-resolution pixels; calculating time-phase change values of various pixels according to the extraction result and abundance of the edge region; calculating a time prediction value and a space prediction value; according to the earth surface homogeneity degree, the time prediction value and the space prediction value, combining neighborhood information to distribute a residual value so as to acquire a preliminary fusion image; and utilizing the established optimization model to correct the change pixels contained in the preliminary fusion image to obtain a spatio-temporal data fusion result. According to the method provided by the embodiment of the invention, the applicability of different change detection algorithms in different scenes is comprehensively considered, the overall spectral precision of fusion is improved, more spatial detail information is reserved, and a better spatio-temporal data fusion result can be obtained.

The invention relates to a Web anomaly detection system based on user behaviors. The system performs request anomaly detection based on user request behaviors, and performs user identity anomaly detection based on user browsing behaviors. For the detection of a request anomaly, an RC-HsMM (Rate Control-HsMM) algorithm combining a time window based request rate sudden abnormal change detection algorithm and an improved HsMM algorithm is applied to user request anomaly detection on the basis of analyzing abnormal behaviors generated by an abnormal user in the process of hacking a website. For the detection for a user identity anomaly, user identity anomaly detection is performed by using a nested One-class SVM algorithm according to browsing behavioral habits and behavioral habits in the session of the user.

The embodiment of the invention provides a remote sensing image change detection method and device, and the method comprises the steps of carrying out the geometric registration of an obtained remotesensing image of a front time phase and an obtained remote sensing image of a rear time phase; carrying out multi-scale segmentation on the remote sensing image of the front time phase to obtain a segmented vector diagram, and nesting the vector diagram on the geometrically registered remote sensing image of the front time phase and the geometrically registered remote sensing image of the rear time phase to obtain a to-be-detected pattern spot pair of the front time phase and the rear time phase; utilizing an automatic change detection algorithm to carry out automatic change detection on the pattern spot pairs of the front and rear time phases to be detected, and determining a non-change pattern spot set and a possible change pattern spot set according to a detection result; and clusteringthe possible change pattern spot set, and applying the clustered possible change pattern spot set to manual judgment to determine a change pattern spot. According to the technical scheme, the efficiency of manually extracting change information in surveying and mapping production can be improved.

The invention discloses a remote sensing image change detection method based on CVA (Change Vector Analysis) and sample selection, which comprises the following steps: registering two remote sensing images of different time phases in a detection area; segmenting the remote sensing images at multiple scales to get image disks; integrating the features of all the image disks to get a difference image; gridding the difference image to get multiple difference image blocks, and calculating the standard deviation; generating a sorting curve according to the difference degree from big to small, and extracting the difference image blocks at the position with maximum slope change of the curve and before the position with maximum slope change; acquiring a change threshold based on a Bayesian threshold, and segmenting the difference image in a binary manner according to the change threshold, and getting a change detection result; and outputting the detection result. The method has the following advantages: the multi-dimensional features of an image object are integrated, the ability of different features of the image object in change detection is displayed, and the reliability and application scope of the change detection algorithm are improved; and the change detection algorithm is very stable.

The invention relates to a high-resolution noctilucent remote sensing image automatic change detection method based on feature fusion. The method comprises the following steps: 1), obtaining two-time-phase high-resolution noctilucent remote sensing data before and after a short-time major event in a research region, and carrying out the preprocessing of front and rear time-phase remote sensing images; 2) based on the preprocessed high-resolution noctilucent remote sensing data, extracting a plurality of derivative texture feature images, and superposing the derivative texture feature images toconstruct a multi-band feature image fused with texture features; 3) carrying out change detection on the multiband characteristic image in the step 2) by adopting a multivariate change detection algorithm MAD and an iterative weighting algorithm IR-MAD thereof to obtain a change intensity graph TMAD and a change intensity graph TIR-MAD fusing multiple characteristics; and 4) respectively segmenting the change intensity graphs TMAD and TIR-MAD to obtain respective corresponding binary change detection result graphs. Compared with the prior art, the method has the advantages of being suitablefor satellite-borne high-resolution LJ1-01 noctilucent remote sensing image processing, high in automation degree, high in precision, long in time sequence, large in monitoring range and the like.

The invention provides an intelligent video monitoring system and method based on multi-layer visual processing. The intelligent video monitoring system comprises a change detection layer, a target detection and tracking layer and a target recognition layer, wherein the change detection layer is used for extracting each frame of image of an image buffer zone in a video input source, and detectinga change region by using a change detection algorithm to obtain position information of the change region; the target detection and tracking layer is used for continuously tracking a detected target and outputting a detection and tracking result in conjunction with a detection result of the change detection layer and each frame of image data in the video input source; and the target recognition layer is used for judging whether the target exists or not, recognizing the target through a target recognition algorithm, and outputting relevant information of the target to a target information storage to be saved. Through adoption of the monitoring system and method disclosed by the invention, the bottlenecks of relatively low efficiency, relatively high cost and great calculation amount duringmonitoring in an existing monitoring system are solved. The intelligent video monitoring system and method can be applied to a wider range of scenarios.

The invention relates to a remote sensing image change detection method. The method includes: inputting two multispectral remote sensing images of the same area and different time phases; carrying outimage registration on the two remote sensing images; performing relative radiation normalization correction on the remote sensing image after image registration, determining change pixels of the reartime-phase remote sensing image after radiation normalization correction relative to the front time-phase remote sensing image by using an iterative weighted multivariate change detection algorithm,and forming a change difference graph by using all the change pixels; processing the change difference graph by using a mean shift algorithm to obtain a change detection image; and finally, adopting amorphological corrosion and expansion algorithm to process the change detection image, so as to obtain a final change detection image. According to the method, the problems of complex background information and serious noise interference of different time-phase multispectral images are solved, the threshold selection difficulty is reduced to the maximum extent, and the spiced salt effect in change detection is effectively reduced, so that the detection result is more reliable and more stable.

The invention provides an SAR (synthetic aperture radar) image change detection algorithm based on non-supervision depth network learning. The algorithm includes the steps: 101 starting an SAR image change detection method based on a non-supervision depth nerve network; 102 performing FCM (fuzzy c-mean) joint classification for two registered SAR images of different time phases in the same area to obtain rough change detection results; 103 selecting noiseless points with large possibility to serve as training samples of the depth network according to initial change detection results; 104 inputting sample points to be trained to the designed depth nerve network to be trained; 105 inputting two images to be detected to the trained depth nerve network to obtain a final change detection result map; 106 finishing the SAR image change detection method based on the non-supervision depth nerve network. Ohm= {ohm1 and ohm2}. Construction links of a difference map are avoided, sensitivity of noise is improved to a certain extent, and detection efficiency and detection accuracy are remarkably improved.

The invention relates to a change detection and test method thereof oriented to a heterogeneous sequence image. The method comprises the following steps that: constructing a standardized test target, and scattering the standardized test target into an infrared image; utilizing a Canny operator edge feature extraction algorithm to carry out edge extraction on stable edge features in the infrared image and a visible light image, and outputting the edge graph of the heterogeneous image; carrying out image blocking on the edge graph of the heterogeneous image, and utilizing a change detection algorithm of a continuous edge probability to calculate the change detection atlases of the heterogeneous image; and comparing the obtained change detection atlases, and selecting the change detection atlas with a high resolution capability to carry out mean value calculation to obtain a change detection threshold value. By use of the method, the change detection can be realized for the infrared image and the visible light image with a great gray difference, and the method has a high change detection rate.

The invention discloses a double-time-phase high-resolution remote sensing image change detection algorithm, which relates to the technical field of remote sensing surveying and mapping geographic information, and comprises the steps of A, image acquisition, that is, a satellite photographing unit acquires multi-time remote sensing image data and registers the multi-time remote sensing image data;and B, image preprocessing, that is, vector, grid and PNG format conversion is performed on each IMG source data image. According to the double-time-phase high-resolution remote sensing image changedetection algorithm, multi-scale features can be obtained by combining the features of regions of different sizes, an image space is averagely divided into sub-regions of a certain scale under the driving of the motivation, a self-attention mechanism is introduced into each sub-region, and by utilizing the space-time relationship of an object under the scale and dividing the image into multi-scalesub-regions, multi-scale feature representation can be obtained so as to better adapt to the scale of the target.

The invention discloses a synthetic aperture radar (SAR) image change detection method based on a multi-objective evolutionary algorithm based on decomposition (MOEA / D) and fuzzy clustering. The problem of balancing two objectives, namely detail maintenance and noise removal in SAR image change detection is solved by a multi-objective optimization method. The method comprises the following steps of (1) generating a difference chart by using a logarithm contrast method for an image to be detected; (2) filtering the difference chart to obtain the denoised difference chart; (3) determining two objective functions according to the two objectives of detail maintenance and noise removal, and combining into a multi-objective optimization problem; (4) obtaining a Pareto front end of the multi-objective problem and a corresponding result chart by using the MOEA / D; (5) selecting a suitable change detection result chart from all results according to a requirement. Compared with a condition that only one solution is obtained by other change detection algorithms, the method has the greatest advantage that the method is used for obtaining an optimal solution set, and a user can select a more suitable solution according to the emphasis on the detail maintenance and the noise removal.

The invention discloses an illumination invariance feature extraction-based remote sensing image change detection method, a device, equipment and a computer readable storage medium. The method comprises the steps of carrying out the image super-resolution of a to-be-detected original remote sensing image through employing a pre-selected image super-resolution algorithm based on deep learning, andobtaining a high-resolution remote sensing image; performing illumination invariance feature extraction on the high-resolution remote sensing image by using an illumination invariance feature extraction algorithm based on a position sensitive histogram to obtain an illumination feature map of the high-resolution remote sensing image; and performing change detection on the illumination feature mapby using a pre-selected remote sensing image change detection algorithm to obtain a change detection result map of the original remote sensing image. According to the method, the device, the equipmentand the computer readable storage medium provided by the invention, the influence of great difference on judgment of the same object in change detection due to different illumination is avoided, andthe accuracy of a change detection result is improved.

The invention relates to the technical field of urban construction, in particular to an all-weather remote sensing monitoring automatic analysis method and system. The method comprises the following steps: S1, obtaining an original image: obtaining original remote sensing images of a land in different periods; s2, geometric registration: carrying out geometric correction on the remote sensing images by using the ground control point data, and carrying out registration between the images; s3, radiation correction: eliminating or correcting radiation distortion or distorted images in the images, and carrying out normalization processing on the images; s4, performing change detection: performing change detection on the image by using a change detection algorithm based on a convolutional neural network and transfer learning, and generating a change result; and S5, result analysis: analyzing the change result, and carrying out early warning processing on the land use condition. Automatic change detection of images in different periods is realized, the change detection precision is improved, the number of manual intervention times is reduced, and rapid change detection result generation is realized.

The invention discloses an ADS-B (Automatic Dependent Surveillance-Broadcast) information-combined deep learning-based airport scene moving target identification method, which combines the idea of change detection motion and static judgment and feature extraction, classification and regression modules in target detection, and improves a time histogram-based background modeling method in a change detection algorithm at the same time. A more accurate background image is obtained; and the existing ADS-B technology of the aircraft is adopted to carry out auxiliary training, so that the feature extraction capability is enhanced, and the detection result is improved. And ideas and advantages of a change detection algorithm and a target detection algorithm are integrated.

The invention discloses a change detection method based on high-resolution remote sensing images. Firstly, after necessary preprocessing such as orthorectification, image registration, and histogram matching, the method uses superpixel segmentation and synthesis The remote sensing image is divided into blocks, and the calculation of local features and sample selection are performed in units of superpixels, so as to realize the automatic labeling of the changed or unchanged areas with obvious tendencies in the image; after that, the twin convolution is trained with the labeling results as samples The neural network classifies the image changes, and performs post-processing such as noise reduction and morphological filtering to obtain the final change detection results. Experiments show that on the Gaofen-2 satellite remote sensing image data set, the indicators of the method of the present invention are much better than the traditional change detection algorithm, the Kappa coefficient is increased by 0.3 on average, the average overall error rate is lower than 3.5%, and the detection results have more High practical value.

A method whereby a presenter may capture images directly from a graphics control circuitry during a presentation without the need for actions, such as mouse clicks, which would interfere with the presenter's ability to present information to an audience. The method automatically captures images using a change detection algorithm and stores those images in digital form for future use. The presenter or user of the method may make adjustments to the parameters, via a control interface, used by the algorithm to determine when to capture and store a static image from the graphics control circuitry.

The invention provides a remote sensing image change detection algorithm based on a coupling discrimination feature self-learning network. According to the algorithm, a coupling discrimination featureself-learning network is mainly designed, the network performs feature extraction on a to-be-detected image through two sub-networks, an established coupling training model can obtain a coupling feature space, the discrimination ability of features in the space is enhanced, generation of a more distinct difference graph is facilitated, and the detection accuracy is improved. An accurate change detection result is obtained. The method does not need label data, is not limited by data types, and has a wide application space. The effectiveness of the algorithm is verified on a public data set.

The present invention proposes a method, system and equipment for spatio-temporal fusion of remote sensing image data. Change detection images are obtained by computing low-resolution remote sensing images in two time phases; the edge area of ​​the high-resolution image in the first phase is extracted, and each The abundance corresponding to the number of high-resolution pixels; calculate the time-phase change value of various pixels according to the extraction results and abundance of the edge area; calculate the time prediction value and spatial prediction value; according to the degree of surface homogeneity, time prediction Values ​​and spatial prediction values, combined with neighborhood information to assign residual values ​​to obtain a preliminary fusion image; use the established optimization model to correct the changing pixels contained in the preliminary fusion image, and obtain the spatio-temporal data fusion result. The method described in this embodiment comprehensively considers the applicability of different change detection algorithms in different scenarios, improves the overall spectral accuracy of fusion and retains more spatial detail information, and can obtain better spatio-temporal data fusion results.