222 results about "Image parsing" patented technology

The invention relates to the technical field of image processing, in particular to an intelligent fracture diagnosis system based on image recognition, which comprises an image analysis module, a mode recognition module, a form analysis module, a risk assessment module and an auxiliary decision module. According to the method, skeleton gray level distribution and boundary consistency are analyzed through continuous frames of X-ray images, fracture feature extraction precision and time sequence coherence are improved, key point space distribution, symmetry standards and form proportions are fused, fracture area structured quantitative evaluation is achieved, the form change trend and abnormal offset point screening are combined, and the accuracy of fracture feature extraction is improved. The method enhances abnormal trajectory recognition precision, associates bone mineral density and form offset, calibrates high-risk time periods, improves risk assessment perspectiveness and individual adaptability, dynamically corrects output content according to an inter-frame suggestion change trend and extended feedback, enhances timeliness of diagnosis suggestions and closed-loop feedback quality, and improves risk assessment accuracy. Image evolution, structural geometry and physiological data are integrally fused, and multi-dimensional intelligent judgment of fracture recognition and evaluation is achieved.

The invention discloses a geological structure fracture automatic identification and detection system based on deep learning, and relates to the technical field of geological structure fracture automatic identification and detection, and the geological structure fracture automatic identification and detection system comprises a fracture image analysis module, a distortion intelligent evaluation module, a form adaptive correction module, a correction efficiency evaluation module and an identification data management module. The distortion intelligent evaluation module is used for acquiring imaging parameter information of each region in the crack image in real time, judging the form distortion degree of each region in the crack image under a perspective distortion condition, and dividing each region in the crack image into a low distortion region, a medium distortion region and a high distortion region according to a judgment result; and the form self-adaptive correction module is used for constructing a fracture form correction mechanism according to the division result of each region in the fracture image. According to the invention, the problem of fracture shape distortion caused by the change of a shooting view angle is solved, accurate identification and automatic correction of the fracture are realized, and the detection precision and stability are improved.

The invention belongs to the technical field of image processing and robots, and particularly relates to an image processing method and device and a robot, and the method comprises the steps: obtaining a structured light coding image, a polarization imaging image and a binocular depth image; constructing a multi-dimensional feature vector by fusing decoded contour phase, polarization analysis material characteristics and depth image data; dynamically calibrating the multi-dimensional feature vector according to the ambient light and the pose of the robot, correcting color deviation through an adaptive white balance algorithm, compensating a camera installation error, and obtaining a calibrated descriptor; based on the calibrated descriptor, utilizing an improved Hough transformation and template matching algorithm to identify the object category attitude, and calculating six-degree-of-freedom pose parameters through a triangulation method; and according to the six-degree-of-freedom pose parameters, a collision-free grabbing path is generated in combination with a robot kinematics model and obstacle information. Therefore, the problems of insufficient image analysis capability, single compensation mechanism, lack of path optimization capability and the like in the prior art are solved.

The invention provides a multi-modal visual understanding model based on double-path visual coding, a training method, a reasoning method and equipment. The model comprises a first visual encoder used for extracting global visual features of a natural general image and outputting first image features, and weight freezing of the first visual encoder; the input of the second visual encoder is an image of which the size is adjusted to a preset high resolution, and the second visual encoder is used for extracting local detail information of the image and outputting a second image feature; the feature fusion layer is used for splicing the first image feature and the second image feature in a channel dimension to form a fused visual feature; the linear layer is used for converting the fused visual features into input dimensions required by a large language model; the large language model is used for generating natural language answers based on the fused visual features after dimension conversion and text input. According to the method, a double-path visual coding structure is adopted, the image analysis capability of the multi-modal visual understanding model is improved, and the alignment mode of the visual features and the large language model is optimized.

The invention relates to the technical field of agricultural intelligent monitoring, in particular to a crop growth detection system and method based on machine vision, and the system comprises an image collection module, a morphological feature capture module, a growth trend judgment module, an abnormal region marking module and a state information output module. According to the method, crop images are collected through a multi-angle camera, leaf contours, stem bending and plant spacing are extracted, multi-dimensional modeling of morphology is realized, structural change identification is enhanced, time sequence comparison of key morphological characteristics is realized, identification precision and time efficiency are improved, leaf and stem change trends are continuously analyzed, offset is quantified, and growth abnormity is early warned in advance; health degradation identification is combined with fluctuation area marking, dynamic monitoring is achieved, high-risk positioning is carried out on a time overlapping area, time-space locking is enhanced, precise management is assisted, the process is from image analysis to abnormal focusing, an information reasoning chain is established, and the monitoring precision and response efficiency are remarkably improved.

The invention discloses a grooving machine rigidity dynamic adjusting system and method based on material image recognition, and relates to the technical field of grooving machine rigidity dynamic adjusting, and the method comprises the following steps: obtaining a material surface and subsurface image sequence of a grooving machine processing area; inputting the image sequence into a pre-trained depth vision analysis model, and outputting a material characteristic parameter set in real time; converting the material characteristic parameter set into a target rigidity value based on a rigidity demand mapping model; within a preset regulation and control window period before the cutter makes contact with the target machining area, the initial rigidity adjustment amount is obtained according to the target rigidity value, and a rigidity adjustment instruction is generated and sent to a rigidity adjustment execution mechanism for dynamic adjustment; in the cutting process, fusion of feedback data and an image analysis result is obtained, and the initial rigidity adjustment amount is corrected; dynamic, accurate and predictive active adjustment of the rigidity of the grooving machine is achieved, and therefore the machining adaptability, intelligence and stability of the grooving machine are effectively improved.

The invention provides an intelligent document management method and system, and the method comprises the steps: obtaining an uploaded file set, processing each file through a text analysis algorithm, an image analysis algorithm and a structured data analysis algorithm, generating a first feature set, and storing the first feature set in a temporary database; obtaining a distribution record from the first permission configuration file, writing the distribution record into an audit database through a log audit module, and generating a first operation traceability chain; according to the first operation tracing chain, a storage management module is called to perform fragmentation processing on the file and upload the file to cloud storage, a first storage index is generated, and the metadata table and the archive position are associated; and according to the second storage index, calling a disaster recovery backup algorithm to perform multi-region copying on the file, generating a first backup confirmation signal, and recording the first backup confirmation signal into an operation and maintenance log. According to the system and the method, intelligent classification, sensitive information protection and whole-course traceable management of the files are realized, the automation level and the safety of document management are improved, and a comprehensive solution is provided for document management.

The invention relates to the technical field of identification control, in particular to a fastener heat treatment integrated digital control system which comprises a process sequence number setting module, a graph code analysis module, a parameter instruction generation module, a temperature zone driving judgment module and a state autonomous response module. According to the method, the fastener category and the thermal process section sequence number are subjected to combined coding and embedded into the graph code to construct the structured recognition pattern, clear identification and automatic recognition of the thermal process section sequence can be achieved, the structured sequence number is automatically extracted in combination with graph code image analysis, and the section sequence index result is generated; an instruction index relation is established through multi-section target parameter aggregation to form a structural body array, a response marking system is established in combination with temperature real-time acquisition and tolerance zone comparison, the abnormal deviation state in the processing process can be dynamically recognized, and autonomous cooling response is implemented based on the temperature deviation frequency; and the process continuity of system operation, the accuracy of a control instruction and the stability of temperature zone execution are comprehensively improved.

The invention discloses a PDF contract file identification method and system, a medium and a program product, and relates to the technical field of information identification, and the method comprises the steps: carrying out the image analysis of an obtained PDF contract file through employing a deep learning model, generating an editable text, and synchronously identifying the page layout of the document, outputting structured data including page numbers, text paragraphs and coordinate information; performing multi-dimensional feature matching on the text paragraphs according to an adaptive semantic analysis algorithm, a preset contract template library and a dynamic keyword library, and positioning contract core element information; performing entity relationship verification on the content of the contract core element information, correcting an extraction error through dependency syntactic analysis and semantic role labeling, and establishing a contract element data set containing a confidence coefficient weight; and performing serialized packaging on the verified contract element data set to generate a contract element list conforming to electronic signature authentication. According to the invention, the processing efficiency and recognition precision of the PDF contract file are improved.

The invention provides an automatic GUI (Graphical User Interface) test method based on visual understanding and large model service awareness, and aims to solve the problems of low test coverage rate and insufficient test depth in the traditional GUI test technology. According to the technology, scene division and business logic of the application page to be tested are deeply understood by combining a multi-modal large model and a control identification technology, and efficient and comprehensive test cases are automatically generated by taking scene business logic as guidance. According to the technical scheme, the system comprises a GUI image analysis module, a page scene division module and a scene-guided test generation module. The GUI image analysis module is used for realizing accurate identification and positioning of a page control; the page scene division module is used for understanding page information by using a large model to realize scene division; the scene-guided test generation module generates a test decision and executes a test based on scene understanding. The method is characterized in that a multi-modal large model is combined with a test generation technology based on scene information for the first time, direct input of a GUI image and automatic test decision generation are realized, a vision and reasoning process is opened, a user thinking mode is simulated to make a decision, and the coverage rate and depth of a GUI test are effectively improved.

The invention discloses a unified streaming processing method, system and device for multi-modal AI interactive content, a medium and a program product, and the method comprises the steps: receiving a request which is sent by a user and comprises an image and a text, and constructing a request context; extracting image features; carrying out image analysis and outputting an image description text; performing intention recognition and outputting an intention; multi-stage reasoning is carried out, and thinking content fragments are generated in a streaming mode; buffering and releasing thinking content fragments; generating text content fragments based on multi-stage reasoning; buffering and releasing the text content fragments; performing recommendation triggering based on the image description text and the intention; recommending content fragments; inserting a recommended position, buffering and issuing; and process event management: issuing event notifications when the beginning, any step fails and the end. The method is a universal method capable of transmitting different types of outputs in a single ordered stream, the analysis cost can be reduced, and the interaction experience and expansibility are improved.

The invention discloses a method for generating an automatic test script of a vehicle-mounted infotainment device and related equipment, and relates to the technical field of automatic test of the vehicle-mounted infotainment device, the method comprises the following steps: generating a control attribute set corresponding to each user interface based on image analysis data and control tree data of each user interface of the vehicle-mounted infotainment device; based on the control attribute set corresponding to each user interface and the hierarchical relationship between the user interfaces, constructing a multi-layer page control knowledge base; and inputting the multi-layer page control knowledge base into a preset large language model to obtain an automatic test script of the vehicle machine output by the large language model. The invention provides a method for generating an automatic test script of an in-vehicle infotainment device and related equipment, the test script is generated through automatic analysis of a large language model according to image analysis data and control tree data of each user interface of the in-vehicle infotainment device, and then the development efficiency of the test script can be improved at least to a certain extent.

The invention provides a quantitative detection method and a quantitative detection system for mesoscopic defects of a soil-rock aggregate based on CT image multi-threshold segmentation, and belongs to the technical field of nondestructive testing and digital image processing of geotechnical engineering materials. The method comprises the following steps: firstly, carrying out multi-threshold segmentation on a CT image by adopting a fuzzy C-means clustering algorithm, and analyzing the image into four types of regions including a background region, a block stone region, a matrix soil body region and a mesoscopic defect region according to gray features; then, based on the heterogeneity characteristic of the material, the mesoscopic defects are divided into matrix defects and interface defects according to defect space distribution; the boundary contour of the block stone is extracted through morphological operation, and intelligent judgment of defect types is realized by combining spatial topology analysis of a defect binary image; and finally quantifying the volume ratio parameters of the two types of defects based on pixel statistics. According to the method, the limitation of a traditional homogeneous material defect analysis method is broken through, and a new technical means is provided for research on damage evolution of the earth-rock aggregate multiphase structure.

The invention discloses a three-dimensional visualization head and neck tumor preoperative imaging system based on deep learning. The method comprises the following steps: image analysis and preprocessing; carrying out image denoising and spatial-temporal feature extraction based on a deep learning convolutional neural network architecture, and introducing a u-Net model to realize image segmentation and semantic annotation; performing three-dimensional reconstruction and surface rendering by adopting a Marking Cubes algorithm, and realizing visual interaction image presentation and personalized operation by combining volume rendering and surface rendering; and data is output through multiple ports, and data viewing and interactive operation of a web terminal and a mobile terminal are supported. The method has the advantages that according to the technical scheme, through fusion of a three-dimensional attention mechanism and multi-phase data, the tumor segmentation Dice coefficient reaches 0.92 + / -0.03, and the method is superior to a traditional threshold segmentation method.

The invention relates to a CSF1R prediction method in brain glioma, a storage medium and a device, and belongs to the technical field of medical information analysis. The method comprises the following steps: preprocessing magnetic resonance image data; extracting a radiomics feature CR of the magnetic resonance image data; performing multi-angle and multi-slice 2D image analysis and processing on the magnetic resonance image data, inputting the processed data into a convolutional neural network for training, and extracting a deep learning feature DL of the last convolutional layer; fusing the radiomics feature CR and the deep learning feature DL to obtain a feature S; and inputting the feature S into a pre-trained machine learning ensemble learning strategy, performing CSF1R label prediction, and finally obtaining a CSF1R prediction result. According to the method, the end-to-end automatic process is realized, the advantages of multi-modal data, radiomics and deep learning are combined, the prediction accuracy and efficiency are remarkably improved, and powerful support is provided for diagnosis and treatment of brain glioma.

The invention discloses a multi-view athlete collision detection penalty auxiliary method for a Kabadi competition, which comprises the following steps: acquiring videos of different views of the Kabadi competition, respectively carrying out target detection, and generating a detection frame of an athlete; performing three-dimensional reconstruction on the position of the athlete according to the detection frame data of the multiple visual angles at the same moment, and analyzing to obtain space-time behavior data of the athlete; for each view angle, according to the athlete detection frame and the spatiotemporal behavior data, detecting whether potential collision occurs in the current frame, and if potential collision occurs in multiple continuous frames, marking the current frame as an event frame cluster; performing posture detection on the athletes in the event frame cluster, calculating the nearest distance between joint points of an attacking party and a defending party, selecting a frame with the minimum comprehensive distance as a penalty key frame, and marking the frame as a collision event; and analyzing and converting the space-time behavior data and the collision event of the athlete and a reconstructed image generated by a plug flow rear end into a live broadcast picture, state information and event prompt for a referee to understand and outputting the live broadcast picture, the state information and the event prompt.

The invention relates to the technical field of computers, and discloses a document format conversion method and device for large model training, and the method comprises the steps: carrying out the classification of a PDF document based on a plurality of classification indexes, and obtaining a document type; when the document type is an image type, performing image conversion and preprocessing to obtain a preprocessed page image; analyzing the preprocessed page image to obtain multi-modal content, and processing the multi-modal content to obtain corresponding processing results; performing content reconstruction and optimization on the PDF document to obtain a first intermediate document; and performing content rearrangement on the first intermediate document to obtain a Markdown document. According to the method, the PDF documents are accurately classified by integrating multiple dimensions, different types of documents are subjected to differentiated format conversion, the format conversion efficiency and accuracy are improved, the generated Markdown document conforms to the original PDF document, semantic coherence and format standardization are achieved, the large model input requirement is met, and therefore the model training effect can be improved.

The invention provides a CAD drawing watching method and device supporting cross-platform font configuration, equipment and a medium, and belongs to the technical field of data processing. The CAD drawing reading method supporting the cross-platform font configuration comprises the following steps: receiving a drawing reading request sent by a front-end operating system; according to an image analysis algorithm, a CAD drawing corresponding to the drawing reading request is analyzed, drawing analysis information is obtained, and the drawing analysis information comprises text information; mapping fonts corresponding to the character information are selected from a built-in font library of a back-end server according to applicable operation platforms of the CAD drawing, and the built-in font library comprises mapping fonts applicable to different operation platforms respectively; and using the mapping font to render a CAD drawing, and sending a reading file corresponding to the CAD drawing to a front-end operating system. The problems of font missing, format incompatibility, rendering incompleteness and drawing information loss or deformation caused by platform incompatibility fonts in the drawing viewing process in the prior art can be solved.

The invention discloses an underground pipeline distribution detection method and system based on a ground penetrating radar, and relates to the field of underground pipe network engineering. The method comprises the steps of performing orthogonal grid division on a to-be-detected interval based on grid preset parameters, performing GPR orthogonal scanning on a grid detection area, and determining an initial trend of an underground pipeline according to a first vertex and a second vertex; adjusting the initial trend of the underground pipeline based on the initial vertical scanning direction; on the basis of the adjusted direction of the underground pipeline, continuing to perform GPR scanning on the grid detection area until scanning of the grid detection area is completed, and inputting obtained radar original data into the image detection model to obtain a first type of parameter value corresponding to the underground pipeline; and inputting the first type parameter value and the second type parameter value into a pipe diameter prediction model to obtain a pipe diameter value. According to the invention, through linkage of the three modules, intelligent and efficient cooperative one-stop service of detection image analysis, detection path planning and detection remote control can be realized.

The invention provides an image analysis processing method and system for a law enforcement recorder, and belongs to the technical field of maintenance management, and the method specifically comprises the steps: carrying out the correlation of the deviation condition of a training course affecting a person in a law enforcement video of an untrained person, and the data of the law enforcement video as a basis; and determining an analysis processing scheme of the law enforcement video associated with the influencing personnel, determining an identification result of the illegal behavior associated with the influencing personnel by using the analysis processing scheme, and determining the law enforcement video associated with the influencing personnel by using the matching condition of the identification result and the law enforcement personnel having the training course of the law enforcement video of the untrained personnel. According to the method, the analysis processing sequence of the law enforcement videos of the untrained personnel is determined, and whether the residual law enforcement videos of the untrained personnel need to be continuously analyzed or not is determined based on the matching condition of the real-time analysis processing result and the law enforcement videos of the untrained personnel which are not subjected to analysis processing, so that the image analysis processing efficiency is improved.

The invention provides a piece counting method applied to a sewing machine, and the method comprises the steps: determining the movement distance of a sewn object when a position sensor detects the movement of the sewn object; if the moving distance of the sewing object meets a moving distance threshold value, acquiring a user hand action image recorded by an image sensor within a set time after the sewing object moves, and analyzing the user hand action image to obtain a hand action sequence; if it is confirmed that the hand action sequence contains the piece counting symbolic actions or the number of times of the hand actions contained in the hand action sequence is lower than the change threshold value, piece counting accumulation operation of the sewn object is executed. According to the method, the distance sensor is combined with the hand action recognition for automatic piece counting, manual piece counting is not needed, and the piece counting accuracy can be effectively improved. The invention further provides a piece counting system applied to the sewing machine, a computer readable storage medium and electronic equipment which have the above beneficial effects.

The invention discloses a cable tunnel multi-source inspection sensing method and system, a medium and a product. The cable tunnel multi-source inspection sensing method comprises the following steps: acquiring a cable tunnel image acquired by a camera of an inspection robot and a point cloud acquired by a laser radar; performing image analysis on the cable tunnel image to obtain a 2D positioning rectangular frame of the cable equipment; the 2D positioning rectangular frame of the cable equipment is converted into a laser radar coordinate system to be associated with the point cloud data; and the 2D positioning rectangular frame of the cable equipment is converted to the point cloud in the area in the laser radar coordinate system, and a 3D target detection model which completes training in advance is utilized to carry out target detection of cable equipment defects. The invention aims to realize real-time monitoring and accurate sensing of a cable tunnel environment by fusing various sensors and an image recognition technology, reduce the labor intensity of inspection personnel and improve the inspection efficiency.

The invention relates to a material tray information input method, system and device and a storage medium, and relates to the field of material trays. Each material disc comprises an information area engraved with material disc information and a slot area used for containing chips. The method comprises the steps that material disc images of a plurality of stacked material discs in an overlook state are collected; segmenting the material tray image to obtain an information image corresponding to the information area; segmenting the material tray image to obtain a chip total image corresponding to the slot position area; analyzing the information image to obtain material tray information; analyzing the chip total image to obtain chip information; and inputting material tray information and chip information, and respectively associating and storing the information image and the chip total image corresponding to the material tray information and the chip information. The technical effect of the invention is that the accuracy of information input of the material tray is improved.

According to the LCD image analysis equipment, the acquisition module is designed to comprise a plurality of different types of interfaces, so that data of different types of screen interfaces of the screen driving board can be captured; and the analysis module is designed to analyze the interface data sent by the acquisition module to obtain image data, and the image data is sent to the display module to be displayed. Therefore, due to the fact that the LCD image analysis equipment is provided with the interfaces of different types, data of the interfaces of different types can be captured, and the application range is wide; according to the LCD image analysis device, the interface data is analyzed through the analysis module to obtain the image data, and the image data is displayed through the display module, so that the LCD image analysis device not only has a data acquisition and capture function, but also has an image analysis function, and the technical problem that a device with a data acquisition function and an image analysis function is lacked in the prior art is solved.

The invention relates to the field of engineering management, and particularly discloses a quality management system and method for engineering construction, and the method comprises the steps: firstly carrying out the preprocessing of a BIM model, and constructing a reference frame integrating visual features and spatial geometric information; when a problem image shot by a mobile terminal at a construction site is received, a computer vision and space calculation technology is introduced, a two-dimensional field image is analyzed into an accurate pose of a shooting camera in a three-dimensional BIM space, and sight line tracing is automatically performed in a BIM digital model based on the accurate pose. And intelligently inferring and retrieving candidate components which are highly matched with the perspective of the problem. And finally, on-site unstructured problem data and the unique identifier of the BIM component are firmly bound through lightweight interaction confirmation with the user, and a standardized structured problem record is obtained. Therefore, the problems of low manual positioning efficiency and unstable association can be solved, and the whole-process intelligent closed loop of quality problem data from acquisition to structured storage is realized.

The invention discloses a synthetic aperture radar (SAR) image colorization system, which comprises a data acquisition module, a model training module and a colorized image generation module, the data acquisition module is used for acquiring a multi-modal data set; the model training module is used for training a noise prediction model by using the multi-modal data set, and the noise prediction module is obtained by constructing a cross attention mechanism network; and the colorized image generation module is used for inputting a to-be-predicted image into the noise prediction model to obtain a color optical image. The method is suitable for disaster monitoring, landform investigation and other scenes needing high-precision SAR image analysis.

Provide a system. , , 【Solution means】 Data acquisition means for a person image in a user terminal, Selection means for a desired impression for the person image, Communication means for transmitting the acquired person image to a server, In the server, analysis means for performing image analysis, Proposal means for proposing clothing and posing based on the analysis result, Generation means for generating a modified image generated based on the proposal, Means for transmitting the modified image to the user terminal, Storage means for storing the modified image selected by the user, A system including.

The invention discloses a document image analysis method and device, electronic equipment and a storage medium. The document image analysis method comprises the following steps: encoding a target document image by using an encoder to obtain target global feature map information of the target document image; and utilizing a joint decoder to perform joint decoding of space and content according to the target global feature map information to obtain a target joint sequence of each target text instance in the target document image, the target joint sequence comprising target space information and target content information of the target text instance. In the embodiment of the invention, the joint decoding of the space and the content can be simultaneously completed by utilizing the joint decoder, the target space information and the target content information of each target text instance in the target document image are obtained, the relevance of the space and the content is enhanced by the joint decoding, the unified processing of multiple tasks is realized, and the processing efficiency is improved. Therefore, the efficiency and accuracy of document image analysis can be improved.

The application discloses a synthetic aperture radar (SAR) image colorization system, comprising a data acquisition module, a model training module and a colorized image generation module; the data acquisition module is used for acquiring a multi-modal data set; the model training module is used for training a noise prediction model by using the multi-modal data set, wherein the noise prediction model is obtained by constructing a cross-attention mechanism network; and the colorized image generation module is used for inputting a to-be-predicted image into the noise prediction model to obtain a color optical image. The application is suitable for disaster monitoring, landform surveying and other scenes requiring high-precision SAR image analysis.

The invention relates to the field of bridge impact data analysis, and discloses a post-impact image recognition and damage quantification system for a bridge structure, which comprises a bridge recognition and analysis module. A dynamic image acquisition module; a model establishing module; a dynamic image analysis module; a damage data evaluation module; an evaluation data quantification module; a bridge damage result analysis module; a bridge repairing module; the invention aims to provide the image recognition and damage quantification system for the bridge structure after collision, and solves the problems that in the prior art, the damage condition of the impacted and damaged bridge is incomprehensively analyzed, so that the bridge repair effect is poor, and the image data before the bridge is impacted is not combined for analysis.