17 results about "Connected-component labeling" patented technology

The present application relates to a method for identifying a type of a shale laminaset based on electrical imaging logging, includes loading and decoding electrical imaging data, and preprocessing the electrical imaging data to generate electrical imaging polar plate image data. An image center cluster is initialized by K-means++ algorithm, finding a new cluster center by iteration according to a mean value after calculating a Euclidean distance, and completing clustering after a threshold or a limit of times is reached. N clustered values are binarized to obtain an end-to-end connected matrix conforming to a principle of wellbore imaging, then a circularly connected component labeling algorithm is carried out, and sinusoidal bedding interface fitting is performed after extracting pixels of laminae by a bedding clustering algorithm. The laminae of a shale gas reservoir is then divided based on electrical imaging logging clustering response characteristic value ranges for different laminasets.

This invention provides a region-based noise reduction method, comprising: S1: Image region labeling: converting a color image into a grayscale image, obtaining a grayscale gradient image, and binarizing it; S2: Image dimension segmentation: dividing the image into different feature regions and labeling them using connected component labeling and edge detection algorithms, and dividing the image into different dimension images according to the labels; S3: Image denoising: denoising using different denoising methods according to the features of each dimension image; S4: Image fusion: obtaining a high-quality image based on the image region features, performing alpha fusion processing between the high-quality image and the corresponding dimension image; finally, superimposing and fusing all dimension-fused denoised images to obtain a denoised grayscale image, which is then converted into a color image for output. The method proposed in this application can achieve better noise suppression while sacrificing less sharpness. It can distinguish different regions of the image by block and perform targeted denoising processing according to the characteristics of different regions.

This invention relates to the field of image recognition technology, specifically to a method for evaluating aggregate skeleton structures based on image processing. The method includes acquiring cross-sectional images of aggregates using an industrial camera; generating standard cross-sectional images using 9-neighborhood convolution and quintic nonlinear mapping; constructing a dual-path semantic segmentation network, including global semantics and local detail paths, to generate binary images; executing connected component labeling and boundary tracking algorithms to calculate the average number of contact points; generating skeleton ratio through pixel integration; and constructing an index-road performance correlation model using uniformity statistically generated by polar coordinate region division to screen aggregate skeleton structures that meet performance requirements. This invention effectively solves the problems of edge artifacts and subjective interference, realizing the scientific transformation of aggregate microstructure to macroscopic mechanical properties, and enabling objective evaluation of pavement material stability.

The invention discloses a window cleaning robot stain identification method and system based on image contrast. The method comprises the steps that S1, a glass surface image is acquired and preprocessed; s2, calculating a reflection intensity index according to the pixel gray scale and the gradient magnitude, and constructing a dynamic weight matrix; s3, constructing the preprocessed image into an observation matrix, and setting a low-rank matrix and sparse matrixes X1 and X2; s4, performing singular value decomposition and thresholding iteration updating on the low-rank matrix; s5, performing total variation constraint update on the sparse matrix X1, and performing L1 norm constraint update on the sparse matrix X2; s6, constructing a contrast diagram according to the difference value of the observation matrix and the low-rank matrix, and marking the positions of the stains with a sparse matrix X2 to form a stain contrast response diagram; and S7, performing threshold segmentation and connected domain marking on the response diagram to obtain a blot area mask. According to the invention, high-precision stain identification of the window cleaning robot under a strong reflection condition is realized.

The application discloses a connected component marking method and device, electronic equipment and storage medium. The connected component marking method comprises the following steps: for the obtained image data, a corresponding row run index table and a run coding table are established; the run index table is scanned from top to bottom, and runs are processed row by row; the run coding table is traversed, and whether the current traversed run is adjacent to the first run of the last row is judged; if the run is adjacent to the first run of the last row, the connected component mark of the first run is assigned to the run, the equivalent relationship between the connected component mark of the current run and the connected component mark of other runs adjacent to the run after the first run in the last row is established, and the equivalence class table is updated; the run coding table is traversed again, if the connected component mark of the current traversed run is 0, a new connected component mark is established for the run, otherwise, the connected component mark of the run is updated by using the mark in the equivalence class table. According to the technical scheme provided by the application, the connected component marking method based on run coding is more efficient and occupies less resources.

This invention discloses a GPU-based neuronal skeleton extraction method based on topology-simplified trunk growth, relating to the fields of neuroinformatics and 3D biomedical image processing. The method includes: threshold filtering and denoising of the input volume data, and retaining the largest connected component based on 3D connected component labeling to remove isolated noise; given the cell body center coordinates, determining the cell body region through spherical expansion and removing it from the foreground to highlight neural protrusions such as dendrites and axons; after obtaining the coarse skeleton, using GPU to perform topology simplification, and constructing a path scoring function combining path length and average normalized brightness, iteratively selecting and retaining the trunk path with the best score, enhancing trunk continuity and suppressing pseudo-branches caused by noise and breaks. This invention can achieve robust skeleton extraction in 3D data with noise interference and uneven brightness, and is suitable for applications such as neuronal morphology reconstruction and quantitative analysis.

The invention provides an adhesion segmentation method and device based on neighborhood feature iteration, equipment and a medium, and belongs to the technical field of adhesion image segmentation. Performing background segmentation on the original image through local adaptive threshold iteration, and performing connected domain marking processing on the first image after background segmentation to obtain a second image; extracting an adhesion target sample in the second image through an area threshold value to obtain an adhesion target sample image; extracting an adhesion target contour in the adhesion target sample graph and carrying out eight-neighborhood feature pattern detection to extract candidate feature segmentation points; and segmenting the original image based on the candidate feature segmentation points. Therefore, according to the adhesion segmentation processing method based on neighborhood feature iteration, the extraction of the wheat pollen in the culture dish and the adhesion segmentation processing can be realized in a complex environment; the method comprises the following steps: extracting a wheat pollen image feature region, and segmenting wheat pollen from a background and mutual adhesion;

This application discloses a dual-channel image acquisition and processing method and a local on-orbit programming method based on a System-on-Chip (SOC). A first SOC processing unit acquires visible light images, and a second SOC processing unit acquires short-wave infrared images; both undergo preprocessing, target extraction, and JPEG compression. Both processing units utilize heterogeneous SOC chips with integrated RISC-V CPUs and FPGAs, and are independently equipped with DDR3 and SPI NOR Flash. The FPGA integrates connected component labeling and centroid extraction modules, employing a single-scan algorithm for parallel processing and directly outputting centroid coordinates at the end of a single frame scan. Local on-orbit programming uses function-level BIN splitting and CRC comparison, only programming the change modules. JPEG compression adjusts image parameters through configuration registers and automatically reloads the quantization table. This application achieves parallel acquisition of visible and infrared light, real-time centroid extraction, incremental programming, and flexible compression, suitable for aerospace attitude control and remote sensing imaging.

The application relates to the technical field of image compression transmission, and discloses a satellite image transmission monitoring station image compression transmission method and system for a reservoir, which comprises the following steps: obtaining an original image, removing system errors to obtain a clear image; extracting a waveband edge intensity and a texture gradient, performing a connected domain morphological close operation after double-threshold screening, and determining a core target region boundary; expanding an adjacent pixel outward by counting a boundary pixel histogram to obtain an expansion region; calculating a saliency map based on the region, screening high saliency pixels, performing connected component marking, and generating a protection area mask; performing wavelet transformation on the expansion region, encoding high-fidelity data according to a mask low quantization step, high quantization encoding of a background region to generate compressed data, and fusing the high-fidelity data and the compressed data to obtain a hybrid compressed image; after verification, when a receiving end decodes, core data is losslessly restored, a background region is reconstructed in detail, and a complete image is output. The method can solve the problem of image distortion.

This disclosure provides methods, devices, and systems for image processing. The present implementations more specifically relate to systems and techniques for binary image processing. In some aspects, an image processing system downsamples an image as a grid of binary cells based on a pooling operation. In some implementations, the pooling operation is a max pooling operation. In some other aspects, the image processing system groups a subset of the binary cells into one or more contiguous regions of the grid based on a binary image clustering algorithm. In some implementations, the binary image clustering algorithm is a connected-component labeling (CCL) algorithm. In some other aspects, the image processing system determines a respective boundary for each of the one or more contiguous regions. In some other aspects, the image processing system maps the determined boundaries to the image. In some instances, the image is a binary motion map of an environment.

A method, system, device, and medium for SAR image road extraction based on DeepLabv3+ and adaptive connected component labeling are disclosed. The method includes: acquiring SAR remote sensing images; preprocessing the SAR remote sensing images to form a SAR image road dataset, which is divided into a training set, a validation set, and a test set; performing data augmentation on the training set; inputting the training set into a DeepLabv3+ semantic segmentation network for training; inputting the test set into the trained DeepLabv3+ semantic segmentation network to obtain a test result image; building a road post-processing algorithm framework based on an adaptive connected component labeling algorithm; using the road post-processing algorithm framework to connect road breaks appearing in the test result image to obtain the final road segmentation result; and using the system, device, and medium to implement the method. This invention improves the accuracy of SAR image road segmentation by applying a connected component labeling module, a single-pixel dynamic programming connection module, and an adaptive road widening module to connect breaks in the CNN prediction result image.

This invention discloses a method and system for quantitative analysis of defects in offshore wind turbine blades based on dynamic routing and heterogeneous expert models. The method includes: acquiring original images of the blades using a UAV; performing global noise reduction and motion blur removal using an attention-based image enhancement model; segmenting the enhanced image into non-overlapping image blocks; classifying the defect types of each image block using a routing model and dynamically distributing them to the corresponding expert models; processing discrete and continuous defects respectively using parallel instance segmentation expert models and semantic segmentation expert models; reconstructing the local mask into a global mask image and performing connected component labeling, pixel statistics, and morphological skeleton extraction to calculate the number, area, and length of defects and generate a detection report. This invention can stably acquire high-quality images in complex marine environments and can intelligently and accurately perform automated quantitative analysis of multi-morphological defects.

The invention discloses a shell type transformer fault diagnosis method under weak turn-to-turn short circuit and a related device, and relates to the technical field of transformer fault diagnosis, and the method comprises the steps: collecting operation parameters of a shell type transformer, and carrying out the preprocessing of the operation parameters; performing multi-scale decomposition on the preprocessed operation parameters by adopting wavelet packet decomposition to generate a two-dimensional gray feature map and generate a sample set; carrying out model training and adaptive optimization on the improved connected domain labeling algorithm to obtain an optimized improved connected domain labeling algorithm; identifying the real-time two-dimensional gray feature map to obtain an abnormal connected domain in the two-dimensional gray feature map; and extracting characteristic parameters of the abnormal connected domain, matching the characteristic parameters with a preset fault sample database to determine a fault type and a fault level, and carrying out spatial positioning. According to the invention, the problem that the fault of the shell type transformer under weak turn-to-turn short circuit cannot be accurately diagnosed due to inaccurate fault feature extraction in a complex scene in the prior art is solved.

The invention discloses an object positioning method based on centroid estimation, and provides the object positioning method based on centroid estimation for solving the problem that robot object positioning precision is insufficient in an unstructured scene, and the method is realized through a four-step cooperative technology: firstly reducing a bounding box focusing target area to eliminate background noise; the object information is converted into a gray level image, and contour features are enhanced in a binarization mode; identifying an effective target area through connected component labeling, and calculating a centroid coordinate; finally, a closed loop is formed by combining negation binary image and centroid visualization verification, compared with a traditional machine learning and 3D point cloud reconstruction method, the method is better in performance in complex scenes such as object overlapping and view angle difference, the problem of low positioning precision in an unstructured scene is effectively solved, and the method is suitable for large-scale popularization and application. The method is suitable for high-precision scenes such as industrial detection, automatic driving and humanoid robot operation, and reliable support is provided for flexible and accurate operation of the robot.

The application discloses a kind of real-time stabilizing and de-jittering systems of monitoring video stream supporting hardware acceleration, comprising: video stream acquisition module: original monitoring video stream data is collected, exposure equalization processing is executed to obtain illumination normalization video frame;Gradient construction module: based on illumination normalization video frame, extract the spatial and temporal gradient amplitude distribution matrix through multi-scale Gaussian difference convolution operation, execute eight neighborhood connected component labeling operation to obtain stable feature point candidate set;Heterogeneous decoupling module: based on stable feature point candidate set, generate global motion vector field through optical flow pyramid hierarchical matching calculation, calculate the local spatial standard deviation of each feature point motion vector in global motion vector field, generate global confidence weight factor dynamically based on local spatial standard deviation;Global motion vector field and synchronous acquisition inertial sensor data are fused based on global confidence weight factor by executing Kalman filtering, and optimal camera pose transformation parameter is obtained.

The invention relates to a multi-source remote sensing image-based pit-pond counting method, which comprises the following steps of: performing median synthesis on a radar remote sensing image and an optical remote sensing image of a research area, and identifying the space distribution of a pit-pond water surface for the synthesized image; extracting non-water body pixels from the water surface of the pit pond through edge detection, and assigning the non-water body pixels as land pixels; for the assigned image, marking a complete pit-pond water body by using a connection assembly marking algorithm; screening pond water bodies with sizes exceeding a preset size value in the complete pond water bodies, calculating a pond water body mean value difference for the pond water bodies by using a moving window, marking pixels with mean values exceeding a preset threshold value, and assigning the pixels as land pixels; converting all the land pixels into vector surface data based on the positions of the center lines of the land pixels; the space distribution of the water surface of the pit-pond is segmented by using vector surface data, and independent pit-pond individual plaques are marked and counted. According to the method provided by the invention, rapid and relatively accurate automatic counting of the number of pits and ponds in a large-scale area can be realized.

This invention discloses a weed identification method integrating deep learning and image processing, comprising: acquiring field images; dividing the field images into several grid images; identifying the grid images using a classification neural network model, and labeling each grid image as either crop or background; using color factors in image processing to perform image segmentation, area filtering, and connected component labeling on the green pixels in the two types of grid images, counting the number of connected components in the two types of grid images, and identifying the weed distribution in the grid images based on the number of connected components. In this invention, the neural network model only needs to focus on whether there are crops in the image, and the crop varieties are single or limited. Even if weed species not appearing in the training set are encountered, it does not affect the identification result. Therefore, this invention can effectively reduce the complexity of weed identification and the cost of building the training set images, and can improve the robustness and generalization ability of the model.