341 results about "Accurate segmentation" patented technology

The invention relates to the technical field of deep learning and computer vision, and discloses a histogram interactive attention-based pavement crack segmentation network processing method and system, so as to enhance the edge detail fidelity and improve the crack segmentation precision. The method comprises the steps of image preprocessing, up-sampling, down-sampling, feature fusion and image reconstruction processing. Wherein global feature modeling in the intensity sub-boxes and among the sub-boxes is realized by constructing a histogram interactive attention module (HIA); a double-branch detail enhancement feedforward module (DDEF) is introduced to enhance spatial detail and high-frequency edge information expression; meanwhile, a Fourier jump enhancement module (FFSM) is adopted to jointly refine jump connection features in a spatial domain and a frequency domain. Through the synergistic effect of the modules, the network can realize continuous recovery and structural consistency modeling of a crack boundary in a complex pavement environment, so that the accuracy and the stability of a segmentation result are remarkably improved.

The invention discloses an object three-dimensional reconstruction method, device and system based on deep learning. The reconstruction method comprises the following steps: acquiring a multi-view color image of an object through a controllable image acquisition device; reconstructing a sparse three-dimensional point cloud by using a motion recovery structure method and obtaining a camera pose; initializing parameters of the three-dimensional Gaussian sputtering model based on the sparse point cloud and performing training optimization; a target object semantic segmentation data set is constructed, and a low-rank adaptive technology is adopted to finely segment all models; generating prompts through an open vocabulary detection model at each view angle, obtaining an accurate segmentation mask, and optimizing a three-dimensional segmentation weight by adopting a joint loss function fusing color consistency loss and edge perception loss; and finally outputting the color three-dimensional point cloud of the target object. According to the method, the original image is segmented, so that the influence of the quality of the rendered image is avoided; the segmentation precision of the model in a specific scene is improved through field adaptive fine tuning; and the accuracy of the segmentation boundary is ensured by adopting a double-loss joint optimization mechanism.

The invention provides a wetland ecosystem health evaluation method based on multi-source remote sensing data, and belongs to the technical field of wetland ecosystems, and the method comprises the steps: carrying out the dense dark pixel and Kalman filtering cooperative atmospheric correction of a multi-spectral remote sensing image to obtain a surface reflectance image, and achieving the precise segmentation of a wetland landscape through a graph cut theory, a spectral unmixing model based on an improved Gaussian kernel is utilized to embed physical constraints to invert water quality parameters, a laser radar canopy height priori constraint three-dimensional radiation transmission model is combined to invert vegetation parameters, and a mixed pixel decomposition result is optimized through a Markov random field. A water quality, vegetation and landscape three-dimensional comprehensive health evaluation system is constructed, degradation dominant factors are identified, and the technical problem that it is difficult for multi-source remote sensing data to cooperatively invert wetland ecosystem multi-dimensional health state parameters is solved.

The invention discloses a road surface meteorological state image segmentation method based on a YOLOv8-BRFA model. The method comprises the following steps: acquiring a near-infrared road surface meteorological image data set; designing a dynamic sparse attention mechanism and a hardware perception neural network based on a YOLOv8 network architecture, and constructing a road surface meteorological state segmentation model in combination with an adaptive decoupling detection head; training the road surface meteorological state segmentation model by using a near-infrared road surface meteorological image data set to obtain a trained road surface meteorological state segmentation model; and obtaining a to-be-detected road surface meteorological image, inputting the to-be-detected road surface meteorological image into the trained road surface meteorological state segmentation model, and carrying out road surface meteorological state identification and segmentation to obtain a positioning and segmentation result of the to-be-detected road surface meteorological image. According to the invention, through structure improvement, the calculation complexity is reduced while the precision of the image segmentation model is improved, and real-time and accurate segmentation and detection of the road surface meteorological state under the complex illumination condition are realized.

The invention provides a camouflage object semantic segmentation method and device based on an adaptive candidate strategy, and a medium, and the technical scheme of the invention obtains the category code of a to-be-segmented image through a coarse feature extractor and a classifier, achieves the transmission of semantic category information, and enhances the semantic perception capability of a subsequent segmentation task. Initial prediction of a camouflage object is carried out through a target detector and an adaptive candidate strategy, an optimal target attention box is generated, and robust space guidance is provided for a core segmentation task; according to the method, category coding and a target attention box are combined, multi-source information is fused and reconstructed through a multi-guide feature fusion module and a multi-task perception decoder, and an accurate segmentation result is output.

The invention relates to a crop plot identification method and system based on time sequence vegetation characteristics, and the method comprises the steps: constructing a normalized difference vegetation index time sequence according to a multi-temporal satellite remote sensing image, and generating a crop probability distribution diagram through a crop classification model; performing connected domain analysis on the crop probability distribution map, extracting the center of gravity of an effective connected domain as a forward attention point, scanning the crop probability distribution map by using a sliding window, and generating a reverse attention point; based on the forward attention point and the reverse attention point, segmenting the high-resolution remote sensing base map through a visual segmentation model, generating a candidate mask set, and screening to reserve a mask with the maximum area as a land parcel segmentation mask of the forward attention point; and combining all the plot segmentation masks to generate a crop plot identification graph, and converting the boundary of each plot segmentation mask into a geodetic coordinate sequence to obtain plot vector boundary data. The method improves the recognition accuracy, and achieves the automatic and precise segmentation of the boundary of the land parcel.

The invention discloses an accurate segmentation method and system for an abdominal organ image contour, and relates to the field of medical image processing, and the method comprises the steps: obtaining a to-be-segmented abdominal medical image sequence, CT image data at least containing one phase, and a corresponding Hensi unit matrix; based on an abdominal organ positioning network, obtaining a region of interest containing a target small organ and a preliminary space prior probability graph; a local HU goodness of fit diagram is generated through the dissection focusing path, a boundary response intensity diagram is generated through the boundary enhancement path, and a refined feature diagram is output; inputting the refined feature map into a main body segmentation network; fusing with corresponding hierarchical features of the decoder network, performing up-sampling and reconstruction, and generating an initial segmentation probability graph of the target small organ; and generating a final segmentation result graph and a comprehensive confidence graph. According to the method, the technical problems that in the prior art, due to the fact that inherent physical attributes are not fully mined and utilized, boundary positioning is fuzzy, the image segmentation precision is insufficient, and the confidence coefficient is low are solved.

The invention relates to the technical field of focus labeling, in particular to a medical image focus labeling method and system based on deep learning, and the method comprises the steps: collecting a multi-modal medical image, and carrying out the noise reduction and contrast enhancement preprocessing to obtain a medical enhanced image; segmenting a focus area through a deep convolutional network and generating a focus labeling mask image; and performing post-processing optimization on the mask image, inputting an optimization result into the sub-pixel level annotation generation model, and outputting a high-precision focus annotation coordinate sequence. According to the method, image quality and focus visibility are improved through multi-modal preprocessing, accurate segmentation is realized by using a deep convolutional network, traditional pixel-level precision limitation is broken through in combination with post-processing and a sub-pixel-level model, the problems of noise interference, contrast difference and boundary blur are effectively solved, and focus labeling fineness and spatial positioning precision are remarkably improved.

The invention relates to a breast tumor lesion part ultrasonic image segmentation method based on an SAM-VMama model. The method comprises the steps of obtaining a breast ultrasonic image set and performing preprocessing; the method comprises the following steps: constructing an SAM-VMama model, wherein the SAM-VMama model comprises a first encoder, a second encoder, a prompt encoder and a mask decoder; obtaining a trained model; obtaining a preprocessed breast ultrasound image to be segmented; and obtaining a segmentation result of the breast lesion tumor. According to the method, the local features are extracted through the first extraction, the global features are extracted through the second encoder, the interactive prompt information processed by the prompt encoder is combined, and the precise segmentation mask is generated through the mask decoder; the method not only retains the flexibility of SAM interactive segmentation, but also improves the long-sequence data processing capability and calculation efficiency through a Mamba structure, effectively adapts to the characteristics of uneven gray scale, multiple artifacts and the like of the ultrasonic image, and achieves the balance of high precision and high efficiency.

The invention discloses a battery microstructure three-dimensional reconstruction simulation method and system based on an electrochemical model and computer equipment. The method comprises the following steps: acquiring a lithium ion battery electrode scanning sequence image, and acquiring a three-dimensional voxel model and an electrode grid chart; s2, acquiring key microstructure parameters of the electrode; s3, performing mesh generation on three-phase voxels in the electrode mesh chart to obtain a mesh file for finite element calculation; and S4, comparing the experimental data of the battery with the simulation data of the three-dimensional microstructure electrochemical model, adjusting the precision of the three-dimensional microstructure electrochemical model, and optimizing related parameters. According to the invention, three-phase accurate segmentation is carried out by using filtering and denoising, image enhancement and connectivity and a watershed segmentation algorithm, and a three-dimensional voxel model containing real particle morphology, cracks and pores and an electrode grid chart are reconstructed; the problems that in the prior art, a lithium ion battery microstructure simulation model is large in idealization deviation and inconsistent with a real structure, and consequently performance prediction is inaccurate are solved.

The invention relates to the technical field of image data processing, in particular to a high-voltage switch shell surface coating uniformity evaluation method and system, and the method comprises the steps: collecting a shell surface coating image; dividing the image into a plurality of super-pixel areas by using a super-pixel segmentation algorithm introducing an adaptive distance measurement mechanism; extracting geometric structure features, texture features and adjacent color difference features of the target area; and carrying out nonlinear fusion on the adjacent chromatic aberration and geometric structure characteristics to construct inter-class characteristics, fusing the inter-class characteristics with texture characteristics serving as intra-class characteristics to obtain a comprehensive score, and judging whether the coating is uniform or not according to the comprehensive score. According to the method, the segmentation weight is adjusted in a self-adaptive manner, and the weak chromatic aberration is amplified in a nonlinear manner, so that accurate segmentation of a tiny defect region is realized; geometric irregularity, color mutation and texture anomaly features are integrated, detection of various tiny coating defects is achieved, and the evaluation accuracy is improved.

The invention relates to the technical field of atmospheric environment monitoring, and discloses a sand and dust path monitoring system, which is characterized by comprising a data acquisition module, an image processing module, a feature extraction module, a path fitting module, a path fusion module and a geographic mapping module which are sequentially cascaded, the data acquisition module is used for receiving remote sensing data of a stationary meteorological satellite and generating a multi-time sand and dust monitoring binary image; the image processing module performs edge detection and morphological operation on the sand and dust monitoring binary image to realize accurate segmentation of a sand and dust region; the feature extraction module comprises a minimum enclosing rectangle calculation unit and an edge analysis unit and is used for extracting four types of feature points of the sand and dust area, and the four types of feature points comprise the lower right corner, the upper right corner, the right side line and the lower side line. According to the invention, by fusing edge detection, morphological operation and multi-feature endpoint identification, precise segmentation of a complex dust region and stable capture of motion features are realized, and the problem that feature points are easy to drift in a traditional method is effectively solved.

The invention, which belongs to the technical field of medical image processing and computer vision, discloses a multispectral imaging-assisted post-radiotherapy skin injury early detection system comprising a multispectral data acquisition module, a deep tissue feature extraction module, a three-dimensional lesion segmentation module and a space-time tracking evaluation module. The subcutaneous 2.5 cm depth tissue information is obtained through multispectral imaging at the wave band of 400-1350nm, blood perfusion, melanin concentration, collagen structure and other physiological parameters are inversed based on the radiation transfer theory, a three-dimensional medical image segmentation algorithm is adopted to achieve three-dimensional accurate segmentation of an injury area, a spatio-temporal evolution model is established to predict the injury development trend, and the damage development trend is predicted. The radioactive skin injury can be detected in the subclinical period, the detection time window is advanced by 5.2 days on average, the occurrence rate of severe dermatitis is reduced by 65%, and a basis is provided for clinical timely intervention.

The invention discloses a structural integrity enhanced camouflage target detection method based on graffiti annotation. The method comprises the following steps: collecting and preprocessing an image; a pyramid vision Transform is used as an encoder to extract a multi-level original feature map; constructing a direction context attention module to optimize the original feature map, aggregating different layers of optimized prediction features, and generating a segmentation prediction result of the camouflage target; constructing an asymmetric local structure consistency loss function, and asymmetrically propagating a supervision signal to a neighborhood pixel by using a high-confidence pixel of graffiti annotation so as to enhance boundary consistency and suppress noise; performing end-to-end optimization training on the camouflage target detection model in combination with partial binary cross entropy loss, structural consistency loss and comparison loss; the method achieves the precise segmentation of the camouflage target under the condition of weak supervision, remarkably improves the boundary definition and structural integrity, and is high in robustness and wide in application value.

The invention relates to the technical field of steel structure grain size analysis, in particular to a microscopic image segmentation result-based steel structure grain size analysis method, which comprises the following steps of: performing format conversion and size unification on a segmentation result of a deep learning segmentation model to generate a complete mask, and performing mask optimization on the complete mask; and generating a visual distribution diagram based on the distribution states of the ferrite masks and the pearlite masks in the complete masks, performing superposition and proportional scale labeling on the original microscopic images, and quantifying the distribution characteristics of the ferrite and the pearlite by combining a uniformity index and a spatial correlation index. According to the method, the ferrite and the pearlite in the microscopic image are accurately segmented through the deep learning segmentation model, the visual distribution diagram is generated on the basis, and quantitative analysis is performed on the tissue distribution in combination with the uniformity index and the spatial correlation index, so that quantitative and objective tissue structure characterization can be realized.

The invention relates to the technical field of material nondestructive testing and image analysis, in particular to a topology preserving-based accurate segmentation method and system for a material pore network, and the method comprises the steps: introducing topological structure priori and connectivity constraints into a plurality of links, such as feature extraction, depth model design, training constraint and post-processing structure repair; the real physical morphological characteristics of the material can be fully utilized in the segmentation process, and the accuracy, connectivity and stability of pore structure recognition are fundamentally improved. According to the overall idea of the method, multi-scale texture information, main channel skeleton information and topological feature information of a material image are used as additional priori to be fused into a segmentation model, so that the model does not only depend on pixel-level classification any more, but understands the morphological law of a pore network from the structural level, and thus real recovery of a complex pore structure is achieved.

The invention relates to the technical field of blood vessel image segmentation, and provides a fundus blood vessel image accurate segmentation method based on a U-Net model, aiming at solving the problem that the existing blood vessel segmentation method based on deep learning is difficult to effectively balance global context and local details. In order to solve the problem of inaccurate blood vessel structure segmentation caused by factors such as poor image quality and low contrast ratio, a method of combining a wavelet and a visual state space model and an adaptive dynamic fusion module is provided to improve the integrity of a slender blood vessel and the structural segmentation accuracy of a complex blood vessel network in a fundus blood vessel image. By integrating multiple efficient feature extraction mechanisms, accurate recognition of small blood vessels and the whole blood vessel structure is achieved, the accuracy and robustness of blood vessel segmentation in complex medical images are remarkably improved, and the method can be widely applied to analysis of eye blood vessels such as retinopathy and hypertensive fundus lesion and research of biomarkers related to the blood vessels.

The invention relates to the technical field of medical image processing, in particular to a prostate MRI image segmentation method based on an anatomical structure relationship, and the method comprises the steps: obtaining a to-be-segmented prostate region and a rectum reference region as an anatomical reference from a prostate MRI image; segmenting the prostate region to be segmented based on a fixed anatomical relationship between the prostate and the rectum by taking the rectum reference region as an anatomical reference to obtain an initial prostate segmentation result; identifying a prostate junction partition region in the prostate initial segmentation result and a rectum junction partition region corresponding to the rectum reference region; and based on the boundary features of the rectum junction partition, performing boundary optimization on the prostate junction partition to obtain an accurate prostate segmentation result. According to the invention, the recognition accuracy of the prostate region is improved.

The invention discloses a lightweight polyp segmentation method based on multi-scale differential features and spatial attention. The method comprises the following steps: preprocessing and dividing an intestinal polyp image data set; then, a lightweight class segmentation model LightRGC-SCNet comprising an encoder-decoder structure is constructed; in the encoding stage of the model, a multi-scale differential feature extraction module is adopted to enhance edges and extract multi-scale features, in the decoding stage, a DySample up-sampling module is guided to dynamically fuse the features and complete up-sampling through joint space attention, and on this basis, a semantic-spatial information fusion module is utilized to collaboratively optimize output features; finally, a combined loss function is adopted to train the model, and performance evaluation and segmentation result output are completed on a test set, a hierarchical processing mechanism and an encoder-decoder architecture are combined, the complexity of the model is remarkably reduced, meanwhile, accurate segmentation of the intestinal polyp is achieved, and the method is suitable for clinical application scenes with limited computing resources.

The invention provides a video stream-based intelligent detection method and system for the state of key equipment of a coal conveying trestle. The method comprises the following steps: acquiring a real-time video stream of key equipment of a coal transporting trestle, extracting a current frame from the real-time video stream, and carrying out image preprocessing; performing inference analysis on the preprocessed video frame by using the optimized deep learning model, detecting a coal plough target and obtaining an accurate segmentation region or key point of the coal plough target; based on the detection result of the coal plough, extracting upper boundary feature points of the coal plough, and fitting an upper boundary line of the coal plough; calculating an included angle between the fitted upper boundary line and a preset lower limit reference line, wherein the included angle represents the real-time opening degree of the coal plough; and when the numerical value is continuously in a preset abnormal threshold value range in continuous multiple frames and is kept stable and unchanged, judging that the opening degree of the coal plough is not in place, and generating alarm information for reporting. According to the invention, high-precision and high-concurrency real-time monitoring of the state of the coal plough is realized, and the false alarm rate is effectively reduced.

The invention relates to a porous medium CT image segmentation method, system and device and a medium, and the method comprises the steps: obtaining a rock core sample, and scanning the rock core sample through a CT device to obtain a CT image of the rock core sample; the method comprises the following steps: preprocessing a CT image of a rock core sample, and extracting a main part of gray level distribution in the CT image as an image to be analyzed; performing multi-distribution fitting on the to-be-analyzed image to obtain a multi-distribution fitting image; and performing intelligent threshold segmentation on the multi-distribution fitting image to obtain an image segmentation result containing different pore filler boundaries. According to the method, through gray feature extraction, multi-distribution fitting, intelligent threshold segmentation and result optimization, precise segmentation of pores containing different pore fillers in the rock is realized, the precision and reliability of pore filler analysis are remarkably improved, and the method can be widely applied to the technical field of intelligent identification of porous medium components.

In semantic segmentation of remote sensing images, boundary information of objects tends to be unclear, and semantic information is imbalanced between different layers of the network, resulting in poor segmentation accuracy and robustness.In particular, there is a demand for a highly accurate segmentation method that combines boundary information and semantic information for areas with complex boundary shapes and fine-grained features. [Solution] The present invention provides a semantic segmentation method for remote sensing images that combines boundary guidance and semantic compensation mechanisms, and constructs a boundary-guided semantic compensation network that includes an EfficientNet-B3-based encoder, a Transformer-based decoder, a cross-layer semantic compensation module, and an auxiliary boundary monitoring module. The network extracts hierarchical features from the input remote sensing image, then uses a cross-layer semantic compensation module to combine high-level semantic information with detailed spatial information, and then uses a Transformer to integrate global context to generate semantic segmentation results. Additionally, an auxiliary boundary monitoring module is used during the training process to perform auxiliary monitoring of boundary information. This allows for compensation of semantic information while suppressing loss of boundary information, improving the accuracy and robustness of semantic segmentation of remote sensing images.

The invention relates to the technical field of image processing, and discloses an image segmentation method and device and electronic equipment, and the method comprises the steps: carrying out the multi-layer coding processing of an original image, generating a corresponding coding result, carrying out the deformable convolution residual extraction processing of a first feature map inputted by a current layer, and obtaining a second feature map, and performing attention processing on the second feature map to obtain a corresponding coding result of the current layer. A common convolution extraction module is replaced by a deformable convolution residual extraction module, so that the boundary segmentation precision can be remarkably improved, a more continuous, smoother and more accurate segmentation boundary is generated, zigzag artifacts and boundary missegmentation are remarkably reduced, and the problem that accurate positioning is difficult to realize is solved; and the computing resources are dynamically guided to the region with the highest diagnostic value by cooperatively utilizing the channel and spatial information, so that the problem that the segmentation process is easily influenced by activation of irrelevant regions is solved.

The invention provides an ultrasonic image segmentation method, device and equipment. The method comprises the following steps: acquiring an ultrasonic image to be segmented; inputting the ultrasonic image into a multi-scale expansion convolution module in a multi-task segmentation model to obtain a first feature map; the multi-scale expansion convolution module is used for extracting cross-scale feature information; inputting the first feature map into an efficient channel attention module in a multi-task segmentation model to obtain an attention weight of each channel in the first feature map; obtaining a second feature map according to the first feature map and the attention weight of each channel; and inputting the second feature map into a decoder module in the multi-task segmentation model to obtain a segmentation result of the ultrasonic image. According to the method provided by the embodiment of the invention, accurate segmentation of the ultrasonic image of the intermuscular brachial plexus retardation scene is realized.

The invention relates to a riverway ice condition identification method based on optics-SAR fusion and adaptive segmentation, and belongs to the technical field of remote sensing image processing and application. Riverway ice condition features are extracted from the optical remote sensing image, and a Ka-SAR feature map is extracted from the Ka-SAR image by adopting an improved high-resolution network; carrying out multi-modal and multi-scale feature fusion on the extracted features, carrying out scale specificity feature extraction by adopting a Gaussian pyramid, and then carrying out weighted fusion on the river ice condition features and the Ka-SAR features on each scale based on a scale specificity weight distribution principle; aggregating the multi-scale fusion features into a final fusion feature map by adopting a bottom-up pyramid reconstruction strategy; improved Kuan filtering is used to optimize the fused feature map, and adaptive threshold segmentation is used to realize ice surface and non-ice surface binary classification in the feature map. The method can achieve the precise segmentation of the ice condition region, and improves the recognition precision of the thin ice region.

The invention relates to the technical field of medical image processing, and particularly discloses an optic nerve sheath ultrasonic image analysis system based on deep learning, which comprises the following steps: acquiring an eye dynamic video stream through ultrasonic equipment, and automatically screening key frames to construct an initial image set; performing multi-scale registration on the initial image and a standard anatomical template to generate a reference image; reversely deriving a displacement vector field based on an image space corresponding relation; performing smooth optimization and geometric reconstruction on the displacement field under the elastic mechanical constraint of the biological tissue to generate a corrected image; reinforcing boundary features in the gradient domain and completing accurate segmentation through edge connection and region growth; and finally, performing morphological verification and diameter measurement on the segmentation result, and outputting a structured analysis report.

The application discloses a method for quantitatively analyzing the infection degree of white spots in the spleen of Pseudosciaena crocea based on machine learning, and relates to a threshold segmentation technology of machine learning. 1) Data acquisition: dissect the Pseudosciaena crocea infected with internal visceral white spot disease, use a high-definition camera to shoot the spleen, and acquire pictures of the white spots in the spleen to be quantitatively analyzed; 2) creation of an interactive segmentation GUI interface: used for reading the pictures of the white spots in the spleen of the Pseudosciaena crocea and automatically converting the pictures into binary images; 3) running of a main program: the main program is written by a python code and is used for the GUI interface to display the selected pictures on the interactive segmentation GUI interface, so that the pictures are provided for contrast segmentation in the next step; 4) picture processing: selecting the pictures, drawing a frame to select a region to be segmented, adjusting a segmentation threshold to acquire a relatively accurate segmentation result, and realizing visual interactive segmentation of the white spots in the spleen; and 5) quantitative analysis: realizing the quantitative analysis of the infection degree of the white spots in the spleen of the Pseudosciaena crocea by the proportion of the white spots in the spleen.

The present application relates to the technical field of intelligent corn straw harvesting, and specifically discloses a straw chopping length detection method and system combining a neural network and image processing, the method comprising: acquiring original images with various morphological characteristics; constructing a YOLOv5 neural network model; outputting the boundary box of the straw target and the reference object; and generating a straw region of interest image; performing image processing on the straw region of interest image and extracting the straw edge; determining the pixel area occupied by the straw based on the straw edge; calculating the pixel length of the straw based on the least-enclosing-rectangle fitting; converting the pixel length of the straw into actual physical dimensions based on the reference object calibration; correcting the actual physical dimensions based on the curvature compensation model; and accurately measuring the length of the chopped corn straw based on YOLOv5 and image processing. The straw region of interest is located by YOLOv5, the target is accurately segmented by traditional image processing, and the original image is processed by the length calculation algorithm, so that the length of the straw can be accurately predicted.

The invention relates to the technical field of image processing, in particular to a bacterial colony counting method based on color space clustering and omni-directional gradient scanning, which is based on an asymmetric clustering model of an HSV color space and combines with an optimized K-means algorithm to realize accurate separation of bacterial colonies and complex backgrounds. Then constructing a geometric center identification system, and predicting all possible colony centers as seed points by using Euclidean distance transformation and a dynamic threshold adjustment mechanism; further, through target class screening evaluation of omnidirectional gradient scanning, a target irregular bacterial colony area is screened out; and finally, introducing a watershed distance-edge constraint algorithm to realize accurate segmentation of high-density bacterial colonies. The reliability of the counting result is ensured through mass center coordinate verification and result visualization optimization. According to the method, the recognition capability of semitransparent bacterial colonies and edge fuzzy targets is remarkably improved, the stable counting performance is still kept under the scene of complex illumination and high-density bacterial colonies, the manual intervention requirement is greatly reduced, and the detection efficiency and the automation level are improved.

The present application belongs to the technical field of high-pressure grouting process optimization, and particularly relates to a high-pressure grouting process optimization method for roadway water inrush cracks. The method first improves the Retinex and dark channel priori combination for image preprocessing; based on the attention-enhanced U-Net, the crack is accurately segmented, and the length, width and depth are quantified; the distributed water pressure sensor is arranged in the crack influence area, the Gaussian smoothing, sliding window linear fitting and correlation analysis are used to remove the abnormal measuring points and calculate the weighted average water pressure; taking the crack geometry and effective average water pressure as the input, a double-branch prediction model is constructed to output the preliminary range of slurry-water ratio, grouting pressure and segmented grouting length; further, a neural network prediction model is constructed from the similar working conditions selected from historical samples, and the improved particle swarm optimization is combined to repair the longitudinal wave velocity of the post-rock mass close to the normal value as the target, and the three types of process parameters are optimized to obtain the optimal grouting scheme.