38 results about "Contour analysis" patented technology

The invention relates to the technical field of target detection, in particular to a search and rescue task-oriented unmanned aerial vehicle thermal imaging visual target detection method, which comprises the following steps of: acquiring multiple frames of thermal imaging images of an unmanned aerial vehicle, extracting regional thermal difference characteristics according to a window, marking a non-background region to generate a candidate set, and fitting and reconstructing a suspected thermal target contour. And analyzing the track and the thermal change rate, screening background interference, identifying jump abnormity, positioning the gravity center, and generating a target repositioning signal. According to the method, the heat value range and variance index sequence in the image area is constructed, the thermal anomaly area is judged in combination with the temperature baseline difference, background disturbance comparison is executed in combination with the direction vector of the coordinate trajectory in the multi-frame image and the thermal change parameter, the false detection probability caused by background noise is reduced, and the detection accuracy is improved. The target jump identification is carried out according to the inter-frame heat value and area change rate in linkage with the thermal isoline closure degree, the target discrimination accuracy in a shielding scene is improved, and the robustness of thermal target extraction in a complex search and rescue environment is integrally improved.

The invention provides an AI vision-based coal cinder amount intelligent detection early warning method and system, and the method comprises the steps: synchronously collecting the surface feature information, thermal radiation feature and internal structure information of a slag hopper of a slag conveyor through a dual-spectrum camera, and obtaining a dual-spectrum image; pixel-level coal cinder segmentation is carried out on the double-spectrum image; based on the segmentation result, counting the pixel proportion of the coal cinder area, and converting the pixel proportion into the actual coal cinder total amount proportion; identifying the segmented coal cinder area by adopting connected domain analysis, dynamically adjusting a size judgment threshold value through contour analysis and geometric feature extraction in combination with the bearing capacity of the slag conveyor, and determining whether large coal cinder exists or not; when the actual coal cinder total amount proportion is larger than the preset threshold value or large coal cinder exists, the risk level is generated to trigger the multi-level early warning control action, the actual coal cinder total amount proportion and the large coal cinder are automatically recognized through double-spectrum fusion segmentation, and early warning is given out when abnormity occurs, so that the accuracy and the real-time performance of coal cinder amount detection are effectively improved.

The invention discloses a workpiece contour projection extraction method and system based on ordered point cloud, and is applied to the technical field of three-dimensional precision measurement. The method comprises the following steps of obtaining high-precision ordered point cloud data of a target workpiece and performing preprocessing; geometric analysis is carried out through the three-dimensional visual point cloud model, and section parameters are defined; performing rapid gridding processing on the ordered point cloud data through an efficient triangulation algorithm; calculating to obtain an intersection discrete point set of the cross section and the triangular mesh; and establishing a section plane local coordinate system and implementing two-dimensional projection transformation to obtain the contour projection of the target workpiece. According to the method, high-precision, high-efficiency and automatic extraction of the section contour of the workpiece can be realized, the actual requirements of industrial field online detection and quality control are met, and the method can be applied to scenes such as part size detection, reverse engineering, quality control and contour analysis in industrial manufacturing.

Techniques are provided for implementing liquid tolerance for a touch device. A sensor array of the touch device is scanned to generate response signals. In response to a response signal exceeding a touch threshold, a signal profile analysis is performed. The signal profile analysis includes sorting the response signals according to ascending order of delta positions to generate sorted signals. An extended vector of the sorted signals is formed. The extended vector of the sorted signals is analyzed using a fitted curve. In response to a feature of the fitted curve satisfying a criteria, the response signal is reported as a touch event. In response to the feature not satisfying the criteria, the response signal is rejected as a false touch.

This invention relates to the field of laser processing and automated control technology, and particularly to a laser heating system, including an equipment device and a control device. The equipment device includes a support platform, a transmission mechanism, a laser, and vision components. The control device includes image acquisition, contour analysis, error compensation, coordinate optimization, and pulse sending modules. The image acquisition module drives the vision component to acquire material images and transmits them to the contour analysis module; the contour analysis module calculates deformation parameters and sends them to the error compensation module; the error compensation module combines pre-stored mechanical errors to generate a compensation matrix and sends it to the coordinate optimization module; the coordinate optimization module calculates and outputs the target coordinates to the pulse sending module; the pulse sending module drives the laser component to move and commands it to emit light. This invention solves the positioning deviation caused by hardware errors and changes in material size, improving control accuracy.

The invention discloses a laser cutting path optimization method of laser cutting equipment, and relates to the technical field of cutting path optimization. According to the laser cutting path optimization method of the laser cutting equipment, geometric primitive data of a to-be-cut workpiece is obtained, contour analysis processing is carried out, and a contour geometric data set of the to-be-cut workpiece is extracted; cutting sequence planning processing is conducted on the contour geometric data set of the to-be-cut workpiece based on a genetic algorithm, and an initial cutting path of the to-be-cut workpiece is generated; multi-target collaborative optimization processing is conducted on the initial cutting path of the to-be-cut workpiece, and the optimized cutting path of the to-be-cut workpiece is obtained.According to the method, the optimized cutting path of the to-be-cut workpiece is verified, the optimized verification cutting path of the to-be-cut workpiece is generated, and therefore the machining efficiency is remarkably improved while the machining precision is guaranteed; therefore, the reliability of the cutting process is comprehensively improved.

A testing method and apparatus for unmanned aerial vehicle (UAV) production includes the following steps: S1, Operating condition construction: controlling the power system of the UAV under test to operate to a preset target speed range; S2, Multi-dimensional data synchronous acquisition: generating a synchronous trigger signal based on the real-time speed signal of the power system; using the synchronous trigger signal to control a strobe light source to illuminate the rotating parts, and acquiring a freeze-frame visual image of the rotating parts during the illumination; simultaneously, acquiring acoustic signals during the operation of the power system; S3, Feature extraction: performing contour analysis on the freeze-frame visual image to extract visual feature values ​​reflecting physical deformation; performing spectral analysis on the acoustic signal to extract acoustic signature feature values ​​reflecting the mechanical operating state; S4, Fusion judgment: inputting the visual feature values ​​and acoustic signature feature values ​​into a preset multimodal fusion model to calculate the comprehensive health index of the power system, and determining whether the product is qualified based on the comprehensive health index.

This invention relates to the field of safety monitoring technology, and more particularly to an intelligent monitoring system for highway tunnel construction safety. The system includes a data acquisition module, a contour analysis module, a feature fusion module, and a data compensation module. By acquiring images of the tunnel surface, dust concentration within the tunnel, and water vapor humidity, the system divides the tunnel into several sub-image regions. The system determines the unevenness and texture complexity of each sub-image region, calculates region-sensitive feature values, and uses the tunnel dust concentration and water vapor humidity, combined with the region-sensitive feature values, to determine whether each sub-image region meets preset conditions. The system also calculates data missing characterization values ​​to determine the modeling anomaly tendency of each sub-image region. Based on the modeling anomaly tendency and the data missing characterization values, the system determines the data acquisition method for each sub-image region. This invention improves the accuracy and efficiency of model building through comprehensive analysis, data acquisition, and compensation of the tunnel environment, thereby ensuring construction safety.

The invention discloses a metal wire rope image feature segmentation method, system and device, and relates to the technical field of metal wire rope defect segmentation, and the method comprises the following steps: collecting a wire rope image; analyzing an edge threshold value between the two strands of wire ropes based on the wire rope contours; analyzing and extracting contour edge intersection points based on the wire rope contour; analyzing an initial contour line of the wire rope image, and analyzing an initial boundary between the two strands of wire ropes based on the initial contour line; calibrating the initial boundary to obtain a final boundary, and dividing the wire rope image into different stranded rope areas based on the final boundary; extracting region features of the stranded rope region, analyzing a defect region, and segmenting specific defect points in the defect region; the metal wire rope defect segmentation method is used for solving the problem that the existing metal wire rope defect segmentation technology cannot ensure whether defect segmentation is effective and accurate due to the fact that the description of image features is not detailed enough and the division of regions is not accurate enough.

This invention relates to the field of visual inspection of label bonding quality, and discloses an intelligent inspection system for label bonding quality, comprising: a polarization excitation illumination module for projecting a circularly polarized light field onto a highly reflective curved substrate; a multi-dimensional polarization imaging module for capturing spatially interwoven polarization signals; a geometric contour analysis module for analyzing the geometric contour information of the substrate; a polarization state compensation matrix generation module for determining the local incident angle and generating a compensation matrix according to Fresnel's law of reflection; a polarization distortion correction module for correcting polarization distortion and outputting a full Stokes polarization vector flow; an interface stress field feature extraction module for extracting phase delay features; and a defect location and judgment module for identifying bubble or wrinkle defects. This invention utilizes phase delay features to characterize the surface deformation of the transparent label bonding interface, and combines this with a polarization state compensation mechanism to remove geometric interference from the curved surface, thereby achieving the identification of hidden bonding defects under highly reflective environments.

This invention provides a method and system for detecting minute defects in proton exchange membranes (PEMs) in fuel cells. The method includes acquiring an original image of the PEM to be detected, preprocessing the original image to obtain a preprocessed grayscale image, the preprocessing including at least grayscale conversion and resolution normalization, extracting full-image phase consistency features from the preprocessed grayscale image to generate a phase consistency feature map, wherein the phase consistency feature map characterizes the structural abrupt changes of pixels in the grayscale image, and then performing multi-scale residual enhancement processing based on the preprocessed grayscale image and the phase consistency feature map to generate a defect-enhanced image. Finally, adaptive threshold segmentation and contour analysis are performed on the defect-enhanced image to locate the defect region in the original image. This method can locate the defect region in the original image of the PEM, thus overcoming the problems in the prior art.

The invention provides a method and a device for detecting a moving object. The method comprises the following steps of: performing difference operation on two adjacent frames of images by adopting an inter-frame difference method to obtain a frame difference image; carrying out binarization processing and morphological processing on the frame difference image to obtain a binarization image; performing contour analysis processing on the binarized image to obtain the area of a contour circumscribed rectangular region of the moving object in the binarized image; dividing a contour external rectangular area in the binarized image into large block motion areas or small block motion areas, performing space-to-channel compression processing on the image corresponding to each large block motion area, and performing image splicing processing on the images corresponding to all the small block motion areas to respectively obtain images to be detected; and performing target detection by using the target detection model to obtain the center point position coordinates and category information of the moving object. According to the moving object detection method and device provided by the invention, the calculation amount of the target detection model can be reduced, and the moving object in a complex scene can be rapidly and accurately detected.

This invention relates to the field of image processing technology, and more specifically, to an intelligent contour analysis method for benzimidazolone pigment particles based on a graph search algorithm. The method includes: acquiring a grayscale image using an electron microscope and performing binarization segmentation; constructing a particle distance map based on Euclidean distance transformation; analyzing the curvature features of pixels in the particle distance map using the Hessian matrix to obtain the probability that a pixel in the particle distance map belongs to the core skeleton of the particle; extracting contour depression points based on the closed contour of the particle region; calculating the adhesion penalty intensity of each pixel in the particle distance map; accurately locating seed points by combining the skeleton probability and penalty intensity; and achieving precise cutting of adhered particles by constructing a graph model containing specific edge weights, using the seed points and background as constraints. This invention solves the problem of over-segmentation or under-segmentation caused by severe adhesion of pigment particles, improving the accuracy of contour recognition.

The application belongs to the field of display screen, and provides a light and dark adjusting method and system for mobile phone display screen, which comprises the following steps: pre-processing input LDR video frames by a fast guided filter to remove noise and artifacts generated by compression or sensors; expanding the brightness range of the input LDR video frames based on a global operator of mid-gray value mapping; while expanding the brightness range of the LDR video frames, iteratively removing false contours in the smooth transition area in the LDR video frames by a minimization algorithm based on iterative projection to a convex set; analyzing the LDR video frames without artifacts and their brightness channels to detect highlights and generate an expansion map, analyzing the expansion map and marking the highlight area, and fusing original brightness details by using a guided filter; and re-distributing the brightness of the LDR video frames after brightness expansion to the RGB color channels by using a Mantiuk formula; the application can realize good display effect when converting LDR to HDR.

The invention relates to the technical field of three-dimensional optical measurement, in particular to a terminal three-dimensional profile structured light measurement method, which comprises the following steps of: 1, setting and calibrating a system; 2, designing and projecting a light mode; step 3, image acquisition and preprocessing; 4, phase extraction and unwrapping are carried out; 5, three-dimensional point cloud generation and optimization are carried out, the three-dimensional point cloud of the terminal is generated through the triangulation principle based on the absolute phase diagram and system calibration parameters, then point cloud optimization is carried out, and point cloud optimization comprises outlier removal and surface smoothing; and step 6, contour analysis and output: three-dimensional contour features of the terminal are extracted based on the three-dimensional point cloud, the three-dimensional contour features comprise edge curvature, surface flatness and critical dimension, and a result is output in the form of a three-dimensional model and a dimension report. An optimized optical mode and self-adaptive frequency selection are designed for the micro size and complex surface of the terminal, and the spatial resolution and precision of measurement are improved.

The invention discloses an image preprocessing method for power supply system file entry, and the method comprises the steps: separating a foreground and a background in a photo based on a semantic segmentation network, removing watermarks and stains through morphological operation, and reserving a target region, thereby obtaining a segmented target region; carrying out feature extraction on the segmented target area, extracting image features by adopting BIC, and comparing the image features with a pre-established photo type template through a feature matching algorithm to determine a key information position; according to the determined key information position, adopting an adaptive threshold segmentation algorithm to extract a target area, and accurately positioning an information boundary through edge detection and contour analysis to obtain the extracted target area; and storing the normalized photo and the extracted information in an archive database, and establishing an index and an association relationship for subsequent image retrieval and analysis.

The present application relates to the field of intelligent image processing and computer vision, and particularly relates to a badminton match video line violation intelligent visual detection method, comprising: constructing a badminton match image dataset containing foot contour and field key point annotation; designing and training a DFSNet foot segmentation network, which introduces an attention mechanism in the decoder part to realize accurate segmentation of left and right foot contours; training a key point detection model based on the annotated key points, and generating a bottom line, a service line and a center line binary mask that conforms to the actual white line in the image through secondary refinement; finally, performing pixel-level overlap detection on the foot contour output by the DFSNet and the refined field line mask, and if there are overlapping pixels, it is determined that the foot has violated the line. The present application takes foot contour analysis as the core, combines deep learning feature extraction with geometric prior optimization, realizes automatic and accurate detection of badminton line violation, and can be widely applied to sports event auxiliary judgment and sports analysis.

This invention discloses a visual detection method for small targets on the surface of an air conditioner outdoor unit based on mask-guided perspective correction. The method includes: 1. Acquiring images of the air conditioner outdoor unit using a visual inspection system, segmenting the unit casing using an improved YOLOv11 instance segmentation model to generate a binary mask; 2. Performing contour analysis on the binary mask, extracting the four corner vertices of the connected regions as source point coordinates; 3. Solving the perspective transformation matrix by combining the prior physical dimensions of the target, correcting the tilted view to a standard front view to eliminate geometric distortion; 4. Extracting candidate regions based on a preset configuration file, and using a template matching algorithm to determine the assembly state of small targets such as screws. This invention eliminates perspective distortion caused by tilted camera shooting through mask-guided perspective correction, significantly improving the detection accuracy of small targets on the surface of the air conditioner outdoor unit, and enhancing detection efficiency and accuracy.

This disclosure provides a method for calculating rock mass quality indicators based on borehole images, relating to the field of engineering geological exploration. The method includes: using a trained structural surface recognition model to identify structural surfaces and fracture zones in a target borehole image; the structural surface recognition model is configured to identify structural surfaces and fracture zones in the target borehole image based on the input target borehole image; the structural surface recognition model is trained using a training sample set of structural surfaces and fracture zones expanded by data augmentation and geometric transformation; based on the depth markers in the target borehole image, the pixel resolution is calculated using edge detection and contour analysis algorithms; based on the identification results of structural surfaces and fracture zones and the pixel resolution, the distance between adjacent structural surfaces and the length of fracture zones are calculated; based on the distance between adjacent structural surfaces and the length of fracture zones, the proportion of short rock segments with lengths less than a preset threshold and the proportion of fracture zones are statistically analyzed, and the rock quality indicator RQD is calculated based on the subtraction method.

This invention discloses a deep learning-based fringe analysis method based on physical information, aiming to provide an efficient and accurate 3D contour analysis method. It includes projecting a single high-frequency fringe pattern onto the test scene using a projector, and acquiring a single frame of fringe image using a camera. An initial low-precision wrapping phase is obtained using a deep learning-enhanced Fourier transform contour measurement method, resulting in an initial phase map of the test scene. A lightweight DNN is used for phase analysis, and the Fourier loss function and phase loss function are calculated and backpropagated to output reliable, high-quality phase results for different types of samples. Combining a phase unfolding algorithm and a 3D reconstruction algorithm, the 3D contour of the test scene is reconstructed based on the relationship between the image coordinate system and the camera coordinate system. This invention improves the accuracy and efficiency of fringe analysis and has broad application prospects, especially suitable for precision industrial inspection and 3D reconstruction.

The application provides a three-dimensional high-speed vibration target clear image acquisition method, system and medium, the method comprising: collecting vibration target images based on a high-speed camera, preprocessing the vibration target images, and extracting target contours; analyzing target center points based on the target contours, measuring three-dimensional position information of the target center points at different time nodes based on a laser displacement meter; analyzing plane vibration information and depth vibration information of the target based on the three-dimensional position information of the target center points at different time nodes; fusing and processing the plane vibration information and the depth vibration information to generate three-dimensional vibration state information of the target; adjusting the viewing angle and focal length of the high-speed camera based on the three-dimensional vibration state information of the target to obtain a clear image of the target; and dynamically zooming and adjusting the angle of the high-speed camera to track high-speed vibration of a complex shape target at a high magnification and acquire a clear image.

The application discloses a paperboard warping degree detection method and device, and relates to the technical field of paperboard surface flatness detection.The paperboard warping degree detection method comprises the following steps of image acquisition, image preprocessing, reflection feature extraction, edge detection, morphological operation, contour analysis and warping degree calculation; the image acquisition is performed by means of a light source and a camera located above the paperboard to be detected to acquire image information of the paperboard; the image preprocessing is performed by converting the acquired image information into a gray-scale image, removing noise and then performing histogram equalization to enhance contrast; the reflection feature extraction is performed by calculating the brightness gradient in the horizontal and vertical directions to obtain a reflection change intensity map, and then performing binarization processing to obtain a reflection feature map; the edge detection is performed by adopting a Canny algorithm to extract a basic edge, and then performing a morphological closing operation on the reflection feature map, and then performing a bit or operation on the basic edge to obtain an enhanced edge; and the morphological operation is performed by performing corrosion operations on the enhanced edge by using different structure elements in sequence to obtain an optimized edge image.

The present application relates to the field of intelligent image processing and computer vision, and particularly relates to a badminton match video line violation intelligent visual detection method, comprising: constructing a badminton match image dataset containing foot contour and field key point annotation; designing and training a DFSNet foot segmentation network, which introduces an attention mechanism in the decoder part to realize accurate segmentation of left and right foot contours; training a key point detection model based on the annotated key points, and generating a bottom line, a service line and a center line binary mask that conforms to the actual white line in the image through secondary refinement; finally, performing pixel-level overlap detection on the foot contour output by the DFSNet and the refined field line mask, and if there are overlapping pixels, it is determined that the foot has violated the line. The present application takes foot contour analysis as the core, combines deep learning feature extraction with geometric prior optimization, realizes automatic and accurate detection of badminton line violation, and can be widely applied to sports event auxiliary judgment and sports analysis.

The invention provides a crop growth situation image evaluation method, which comprises the steps of capturing an original image of a crop stalk growth situation, processing the original image by adopting a skeleton extraction algorithm to separate out main structure lines of stalks, and obtaining a stalk skeleton contour; the directions of the base and the top in the bending direction angle are analyzed, the included angle difference value of the bending directions of the base and the top is calculated, when the directions are opposite, the segmentation weight is adjusted to increase the proportion of the top segment, and the initial rating of the phototactic growth tendency is obtained; according to the combination of the transverse distance of the gravity center of the stalk and the segmented weight of the bending curve, the distinguishing score of wind damage inclination and phototactic growth is obtained through evaluation; the bending direction angle, the preliminary rating and the distinguishing score are integrated, a support vector machine algorithm is adopted, three indexes serve as input, feature fusion processing is conducted on the whole stalk situation, and the comprehensive score of the stress resistance is obtained.

The invention provides an OCR (Optical Character Recognition)-based hand-filling table data structured recognition method, which comprises the following steps of: acquiring a scanning image of a hand-filling table, and performing effective table region segmentation on the scanning image to obtain a table image; performing table contour analysis on the table image to obtain a contour cell coordinate frame, and identifying character attributes in the contour cell coordinate frame; deducing a merging cell range based on the contour cell coordinate frame and the character attributes, and extracting a logic cell coordinate; based on the extracted logic cell coordinates, identifying text contents in the corresponding logic cells by adopting a domain self-adaptive CRNN model, and outputting a character sequence; and inputting the logic cell coordinates and the corresponding character sequences into a preset data template to generate a structured table data file. According to the method, the integrity and consistency of the long table data in the digitalization process are ensured, the recognition precision and efficiency of the hand-filling table are improved, and the method is particularly suitable for complex hand-filling tables with incomplete wireframes and irregular formats.

This invention provides a method and apparatus for generating parameters for punched standard parts, relating to the field of parameter generation technology. The method includes: determining the hole boundary of each hole position based on a pre-acquired STL model; performing contour analysis and arc fitting on the hole boundary of each hole position to obtain the hole type, geometric parameters, body center, and normal vector of each hole position; determining a unified selection entry parameter for each hole position based on the hole type and geometric parameters; and determining the standard part parameters of the punched standard part for each hole position based on the unified selection entry parameter, body center, normal vector, and a standard part library. This invention can generate accurate standard part parameters for punched standard parts based on a pre-acquired STL model without relying on existing process models and completed engineering annotations.

The present application relates to target detection technical field, specifically to a kind of unmanned aerial vehicle thermal imaging visual target detection method for search and rescue task, comprising the following steps, obtain unmanned aerial vehicle multiple frames thermal imaging image, extract area thermal difference feature according to window, mark non-background area to generate candidate set, fit reconstruction suspected thermal target contour, analyze trajectory and thermal variation rate and exclude background interference, identify jump anomaly and locate barycenter, generate target relocation signal.The present application, by constructing the thermal value range and variance index sequence in image area and combining temperature baseline difference judgment thermal anomaly area, the direction vector of coordinate trajectory in combination with thermal change parameter executes background disturbance comparison, reduces the false detection probability caused by background noise, according to the thermal value and area change rate linkage thermal isoclines closed degree to carry out target jump identification, improve the target discrimination accuracy under the shielding scene, overall improve the robustness of thermal target extraction in complex search and rescue environment.

The invention discloses a time watermark identification method and system for a machine vision picture, a terminal and a storage medium, and the method comprises the steps: obtaining a to-be-processed machine vision picture of a target machine, carrying out the watermark region positioning of the to-be-processed machine vision picture, obtaining a time watermark region, and storing the time watermark region in the target machine; obtaining a watermark region image according to the time watermark region; preprocessing the watermark area image to obtain a target watermark area image, and performing character recognition on the target watermark area image to obtain a character recognition result; and performing semantic error correction on the character recognition result to obtain a target time character string, and obtaining target time information according to the target time character string. According to the method, the watermark position is dynamically determined through watermark contour analysis and corresponding rules, semantic verification and character error correction are carried out on the character recognition result, the anti-interference capability is improved, and the recognition efficiency and the recognition accuracy are improved.

The invention discloses a Raman temperature and humidity dynamic detection method based on self-adaptive light intensity adjustment, and belongs to the technical field of Raman detection. The method comprises the steps of collecting and processing spectral data to generate a preorder feature contour set, constructing a temperature and humidity parameter detection pair, simulating interference to generate a postorder feature contour set, analyzing contour feature parameters, calculating fluctuation, screening an optimal detection pair and feeding back. The system correspondingly comprises a spectral data acquisition and processing module, a temperature and humidity parameter configuration module, a feature contour analysis module, a fluctuation calculation and screening module and a feedback execution module. By dynamically adapting the optimal temperature and humidity parameter detection pair, the problem of insufficient detection precision caused by light intensity interference and low parameter matching degree in traditional Raman temperature and humidity detection is solved, manual intervention on parameter adjustment is not needed, the detection stability and efficiency are improved, and the method is suitable for various high-precision temperature and humidity detection scenes such as industrial production and environment monitoring.