17 results about "Morphology Method" patented technology

The application provides a remote sensing image building extraction and segmentation result regularization method, which comprises the following steps: acquiring a remote sensing image, wherein the full-color spatial resolution of the remote sensing image is greater than 1 m; cutting the image picture to 512*512 as input data; extracting a building probability map in the image picture by using an improved FCN based on DeepLabV3+; and regularizing the probability map by using a trained generative adversarial network, wherein the trained generative adversarial network comprises a generator and a discriminator, and the generator comprises an encoder and a decoder; the full convolutional neural network is trained by using the generative adversarial method, is used for post-processing of a semantic segmentation result of a building in a remote sensing image, and is used for regularizing the semantic segmentation result; compared with a traditional morphological method, the influence of errors in semantic segmentation on subsequent regularization results is reduced, so that regular building contours are generated.

This specification discloses a method, apparatus, and system for processing three-dimensional tooth images. The method includes: acquiring a 3D original image containing teeth; identifying tooth voxel regions from the 3D original image; selecting crown feature voxels from the tooth voxel regions based on the grayscale information of the crown; determining a crown mask based on the crown feature voxels using morphological methods; and segmenting each tooth from the tooth voxel regions according to the mask to obtain a tooth segmentation result. This method uses morphological methods to determine the crown mask to be compatible with the actual dentition, resulting in high accuracy when segmenting each tooth from the tooth voxel regions according to the mask. This solution is highly adaptable to cases of malocclusion or tooth defects. This solution can segment individual teeth from three-dimensional tooth images fully automatically or with minimal manual assistance, achieving high segmentation efficiency and accuracy.

The invention discloses a method for identifying the purity of a coconut strain. The method comprises the following steps: collecting DNA of a target sample; constructing a DNA library of the target sample, and performing whole genome re-sequencing; filtering data in the DNA library to obtain an impurity-removed DNA sequence; comparing the DNA sequence subjected to impurity removal with a reference genome, carrying out variation detection, generating SNP and INDEL marker files, then carrying out genetic typing, and constructing a variety identification model; comparing and analyzing whole genome SNPs of the two groups, and screening specific sites capable of identifying and distinguishing the two groups in the genome; and inputting the DNA of the target sample into the variety identification model, detecting the SNP specific site of the gene, and judging the variety result to realize the identification of the purity of the coconut variety. The identification accuracy of the identification method can reach 99% and is far higher than that of a traditional morphological method, rapid identification can be conducted in the seedling stage, and the identification period is greatly shortened.

The present disclosure provides a method for processing a cell image, an electronic device, and a storage medium. The method includes: obtaining, based on a morphological method, a target cell nucleus map according to an initial image segmentation result of a target cell image, where the target cell image is an image corresponding to at least one cell; obtaining, based on the morphological method, a target cell region contour map according to the target cell nucleus map and the initial image segmentation result; obtaining, based on the morphological method, a target cell contour map according to the target cell region contour map and the target cell nucleus map; and obtaining a target image segmentation result according to the target cell contour map and the target cell nucleus map.

The invention discloses a cold region concrete freeze-thaw damage detection method and system based on machine vision, and the method comprises the steps: obtaining an RGB image and three-dimensional depth information of the surface of a concrete structure through a high-resolution depth camera carried by an unmanned plane, and carrying out the detection of the freeze-thaw damage of the concrete in combination with the technologies of image preprocessing, large model segmentation, three-dimensional contour reconstruction and the like. Damage features such as aggregate exposure and matrix peeling of concrete under the action of freeze-thaw cycle can be comprehensively reflected, a multi-dimensional parameter system is established to realize visual expression of a damage state, an SAM model and a morphological method are adopted to identify an aggregate area, a depth image is matched to construct a real geometric contour, and the accuracy of the aggregate area is improved. The surface degradation process is more accurately described from two-dimensional and three-dimensional angles, and meanwhile, the freeze-thaw training times are intelligently judged by means of a machine learning model, so that the analysis efficiency and objectivity are improved.

The invention belongs to the technical field of hydroelectric generating set monitoring, and particularly provides a hydroelectric generating set axis trajectory recognition method based on multidimensional features and CFAWOA-SVM, which comprises the following steps: firstly, preprocessing an axis trajectory image by adopting a morphological method, then extracting contour, moment and geometric feature sets of an axis trajectory sample, and combining into a multidimensional feature set; then feature selection and dimension reduction are realized by adopting an RF-Fisher method, and finally, a trained CFAWOA-SVM model is adopted as a classifier to identify an axis trajectory mode. The hydroelectric generating set axis track identification method provided by the invention has relatively high identification precision on the axis track of the hydroelectric generating set, and also has relatively high identification efficiency and relatively good robustness.

Systems and methods provide an antialiasing engine and its related functions that distribute the computational effort and image reconstruction functions to achieve higher visual quality using morphological methods and operations. For example, an antialiasing engine detects one or more edges within a first frame and generates an edge image indicating a respective pixel's proximity to the foreground or the background based on detection of the one or more edges within the first frame. The antialiasing engine then generates an analytical edge by tracing the edge image and encodes the analytical edge into a depth frame within a video stream. The video stream is then transmitted to a client device where the analytical edge is responsively decoded by a client-side antialiasing engine and sampled to generate a final composition integrating the first frame into a second frame using the analytical edge, where the second frame is generated by the client device.

The present application relates to a kind of tower tilt online monitoring method based on remote sensing satellite optical image, comprising: step 1: obtain remote sensing satellite optical image, and carry out segmentation, obtain the image of tower shadow area;Step 2: with morphological method, the image of tower shadow area is enhanced and segmented, adopt a kind of adaptive variable structure element morphological edge detection operator, by Top-Hat transformation and Bottom-Hat transformation, remote sensing satellite optical image is high processing, then again using Canny operator carries out edge detection;Step 3: using Hough straight line detection algorithm extracts the straight line of tower shadow edge in tower shadow image, calculates the midpoint of the coordinates of contour line, obtains the geometric center line of contour.The step 4: establish the geometric relationship model of tower tilt and tower shadow;Step 5: based on step 4, calculate the inclination angle, according to inclination angle, judge tower tilt situation, and output result.

The present application relates to a kind of article volume measurement methods of preventing personnel interference.The article volume measurement method includes S1, obtaining article image, removing background information in image, and carrying out binary processing;S2, find the outline of article, judge whether the outline of article extends to the boundary of field of view, if yes, turn into step S4, if no, turn into step S3;S3, carry out the volume measurement of article, turn into step S7;S4, by morphological method, remove irregular part in the outline of article, if can remove, turn into step S5, if can not remove, turn into step S6;S5, according to the outline of article of removing irregular part, carry out the volume measurement of article, give first-level warning, turn into step S7;S6, give second-level warning, turn into step S7;S7, end.The article volume measurement method of preventing personnel interference proposed in the present application can effectively improve the efficiency of article volume measurement.

Systems and methods provide an antialiasing engine and its related functions that distribute the computational effort and image reconstruction functions to achieve higher visual quality using morphological methods and operations. For example, an antialiasing engine detects one or more edges within a first frame and generates an edge image indicating a respective pixel's proximity to the foreground or the background based on detection of the one or more edges within the first frame. The antialiasing engine then generates an analytical edge by tracing the edge image and encodes the analytical edge into a depth frame within a video stream. The video stream is then transmitted to a client device where the analytical edge is responsively decoded by a client-side antialiasing engine and sampled to generate a final composition integrating the first frame into a second frame using the analytical edge, where the second frame is generated by the client device.

The invention discloses a big data identification method for periodic weighting of a coal face, and belongs to the technical field of coal mine pressure analysis. The method comprises the steps that S1, support resistance data are cleaned based on pressure gradient characteristics, and process noise is eliminated; s2, constructing a nonlinear propulsion model, and realizing accurate mapping from time domain data to a space propulsion degree; s3, the working face is divided into regions and boxes, the group energy of each region is calculated, and a multi-dimensional energy sequence is constructed; s4, extracting a static load baseline by adopting an anti-noise trend separation method based on quantiles, separating load disturbance characteristics, and enhancing by using a morphological method; and S5, constructing an adaptive threshold based on the robust statistics, and performing consistency verification in combination with the geological prior periodic step pitch to realize automatic and accurate identification of the periodic weighting. According to the method, the problems of incomplete data cleaning, distortion of space-time mapping, weak anti-noise capability, dependence on empirical thresholds and the like of a traditional method are solved, and the accuracy, the robustness and the automation level of weighting identification are remarkably improved.

The application discloses a kind of periodic weighting big data identification methods of coal mining face, belong to coal mine strata pressure analysis technical field.The method includes: S1, based on pressure gradient feature cleaning support resistance data, eliminate process noise;S2, construct nonlinear propulsion model, realize the accurate mapping of time domain data to space propulsion degree;S3, working face is divided and is boxed, calculates each regional group energy, constructs multidimensional energy sequence;S4, using the anti-noise trend separation method based on quantile extracts static load baseline, separates load disturbance characteristics, and is enhanced using morphological method;S5, based on robust statistics, construct adaptive threshold, combine geological priori period step distance and carry out consistency check, realize the automatic, accurate identification of periodic weighting.The application solves the problems of traditional method, such as incomplete data cleaning, space-time mapping distortion, weak noise resistance and identification relying on empirical threshold, significantly improves the accuracy, robustness and automation level of weighting identification.

A method and apparatus for early warning of power equipment faults based on infrared image area expansion is disclosed. The method comprises the following steps: acquiring an infrared thermal image sequence of power equipment in a substation using an infrared camera; preprocessing the infrared thermal image sequence using morphological methods to solve for the difference image of the infrared thermal image sequence; performing edge detection on the two-dimensional matrix of the difference image using the Sobel operator to generate a new two-dimensional pixel matrix; scanning the new two-dimensional pixel matrix line by line to obtain a target region; extracting the endpoints and vertices of multiple contour curves from the target region; connecting the multiple contour curves based on the endpoint set and vertex set to obtain a closed curve; and solving for the area of ​​the closed curve; and realizing early warning or alarm of power equipment faults based on the change in the area of ​​the closed curve corresponding to the difference image. This invention reduces the probability of faults being obscured and thus advances the fault detection time.

The present application relates to a kind of structural light projection's luggage three-dimensional version shape degree of fit real-time detection method, to solve the problem of insufficient adaptability to complex geometric boundary identification and process tolerance in traditional fit detection.The core scheme includes: collecting and analyzing the three-dimensional point cloud of luggage surface, calculating the principal curvature direction and change rate to generate curvature distribution map, extracting process boundary by first derivative and morphological method, realizing process constraint mapping by combining knowledge base area semantics, generating point-by-point fit degree weight field by fusing curvature and semantic weight, then combining with measured gap data to obtain weighted deviation, and implementing enhanced correction in boundary narrowband.Region normalization statistics are output, and quantitative fit score is output.Finally, it is automatically compared with tolerance threshold, and the defect precision alarm is realized.The method improves the accuracy of boundary detail identification and the intelligent level of fit detection, adapts to different types of luggage and complex process rules, and realizes efficient and scalable fit quality control.

The application discloses a Chinese environment mathematical formula extraction and recognition method based on a Transformer. The method comprises the following steps: inputting a picture containing a mathematical formula into a mathematical formula mask extraction module in a Chinese environment, and extracting a mask containing the mathematical formula, wherein only the mathematical formula part is reserved and the text part at other positions is removed in the mask; inputting the mask into a mathematical formula picture extraction module, and further repairing and perfecting the mathematical formula position in the mask through a morphological method, and generating a mathematical formula picture according to the repaired mask; inputting the mask into a conversion display module, and obtaining the Latex of the formula through the module, and generating a mathematical formula picture through Katex to display and compare. The method improves the accuracy of formula extraction and formula recognition based on the Transformer technology. Unlike the prior art, the application can automatically infer the formula region of the input picture, and is more convenient to use.

These disclosures relate to a method for detecting and removing personnel interference while measuring volume of an object. An image of an object is obtained, a background information of the image is removed, and the image is converted into a binary image. An outline of the object is detected from the binary image. A determination is made whether the outline extends towards a boundary of the field of view. If the outline of the object does not extend to the boundary of the field of view, the volume of the object is measured. Otherwise, an extending part is removed from the outline using a morphological method, if removable. If the extending part is removed, the volume of the object is measured based on an outline of the object obtained after the extended part is removed, and a class 1 warning is issued. Otherwise, a class 2 warning is issued and the measurement is ended. This approach improves efficiency of the object volume measurement.

This application discloses an image processing method, apparatus, and electronic device, comprising: acquiring a set of discrete contours to be processed in a target image; determining pairs of contours that are adjacent to each other in the set of discrete contours based on preset distance conditions; grouping and aggregating directly or indirectly adjacent contours in the set of discrete contours according to the pairs of adjacent contours to form at least one contour connected group; determining the aggregation boundary region corresponding to each contour connected group; and performing subsequent image processing on the target image based on the aggregation boundary region. This application significantly improves the accuracy and robustness of keyboard character segmentation and localization, laying a solid foundation for subsequent character detection, and avoiding the problem of erroneous adhesion of adjacent characters caused by the fixed kernel size in traditional morphological methods.