282 results about "Corner detection" patented technology

The invention discloses an aluminum profile surface defect automatic detection method and system based on image recognition, and relates to the technical field of image processing and intelligent detection.The method comprises the steps that three-dimensional geometric data of the section of an aluminum profile are obtained through a three-dimensional scanning device, candidate observation angles and geometric feature parameters needed by coverage calculation are extracted, and the three-dimensional geometric data of the section of the aluminum profile are obtained; if the section shape contains a groove or a curved surface structure, marking a space coordinate range corresponding to the area of the shadow region to obtain a section geometric feature vector and a shadow region coordinate set; according to the aluminum profile surface defect automatic detection method and system based on image recognition, all-dimensional dead-corner-free detection of the surface of the aluminum profile is achieved, the method and system can adapt to the production takt of complex section shapes and changes, the accuracy and comprehensiveness of defect detection are improved, and effective technical support is provided for aluminum profile quality control.

The invention discloses a two-stage infrared and visible light image registration method, system and device. The method comprises the following steps of image preprocessing, contour extraction, angular point detection, feature description and matching, affine transformation estimation, multi-scale optical flow estimation, optical flow constraint and loss function, reverse resampling and fusion and error evaluation. According to the invention, rough registration is carried out by using contour angular point features, so that a preliminary alignment result can be quickly obtained; refined alignment is carried out in combination with an unsupervised optical flow network, and sub-pixel-level registration precision is achieved. The contour angular points are based on shape information of an image target, are natural and are not influenced by spectral differences, and the matching stability is enhanced through main direction and angle features. The unsupervised depth optical flow model estimates a pixel displacement field by learning consistency characteristics of an input image, and does not need to depend on annotation data. The combination can effectively eliminate the difference between infrared light and visible light, and improves the robustness and adaptability of registration.

The present invention relates to the field of rock structure measurement, and disclosed are a three-dimensional pore reconstruction method and system based on a rock image, for use in solving the problem that using a machine learning method to reconstruct pores in a rock image requires relatively high mathematical or computer science expertise, involves high labor and economic costs, and is difficult to apply to small-scale rock pore reconstruction projects. The method comprises: step 1: grouping slice images of each rock sample into one set, and performing preprocessing; step 2: performing grayscale processing, histogram equalization and normalization processing on an image; step 3: using a Harris corner detection algorithm to perform corner detection, sorting Harris response values, selecting key points by setting the number of key points, and using a non-maximum suppression method to screen the selected key points; step 4: segmenting the boundaries of pores in the image by means of an adaptive threshold selection method, and extracting pore features; and step 5: using a marching cubes algorithm to construct a three-dimensional pore reconstruction model. The present invention is used for three-dimensional reconstruction of pores in a rock image.

The invention discloses an image processing method and device, equipment and a storage medium, and is applied to the technical field of image processing. According to the scheme, the method comprises the following steps: detecting angular points in a target image to be interpolated based on an angular point detection algorithm to obtain feature points in the target image; for the texture complexity of the neighborhood image of each feature point, selecting a target interpolation algorithm for processing the feature point; and performing interpolation processing on the feature points based on a target interpolation algorithm to generate an interpolation result image. For angular points with complex textures, a high-precision interpolation algorithm is adopted to ensure that image details are finely restored; and for the angular points with slightly simple textures, a low-calculation-amount algorithm is used to reduce resource occupation, so that the limitation that a single algorithm is uniformly adopted for the whole image in the prior art is avoided, and the redundant calculation amount is remarkably reduced while the interpolation precision of the key feature region is ensured and the quality of the whole image is maintained.

The embodiment of the invention provides a pose estimation method and device and electronic equipment, and belongs to the technical field of visual inspection and positioning. The method comprises the following steps: acquiring a target image in real time, wherein the target image is an image of a target object containing a similar two-dimensional code ARuco; performing image preprocessing according to the target image to obtain binary image data; performing mark detection processing according to the binarized image data and a pre-established multi-scale image pyramid so as to identify and position a similar two-dimensional code ARuco mark; performing camera calibration processing according to preset camera parameters to obtain camera calibration data; performing corner detection processing according to the ARuco mark to obtain positioning corner position information; and according to the positioning corner point position information, the preset target three-dimensional space position information and the camera calibration data, carrying out mark attitude calculation processing to obtain target pose estimation information. The method is used for achieving the effect of improving the accuracy of the pose estimation result.

The invention relates to the technical field of camera testing, in particular to a camera triangulation distance measurement detection method. Comprising the following steps: selecting a tested camera, and adjusting the position and angle of the camera; according to the positions of the two targets in the image, setting test areas corresponding to different view field angle pictures, and adopting an area detection mode; angular point detection is carried out on the set area, the angular points and sub-pixel-level angular points are obtained, and checkerboard center angular points are judged and corrected; according to a distortion calibration model, selecting a corresponding function to carry out distortion removal calculation on the angular point to obtain a normalized distortion-removed angular point coordinate, and converting the normalized distortion-removed angular point coordinate into an actual distortion-removed pixel coordinate; calculating the calculation distance of two target angular points of each image, establishing a normalized three-dimensional coordinate, calculating the deviation angle of the left and right angular points by using the three points, and combining the measured parameters to obtain the calculation distance of the two angular points; and calculating a precision error. According to the technical scheme, the accuracy of the internal reference calibration data and the reliability of the distortion model can be effectively verified.

The invention discloses a house type recognition and three-dimensional reconstruction method and system based on grating images, and belongs to the technical field of house type image recognition. The method comprises the steps of 1, collecting and screening high-quality house type images suitable for national conditions of China regions, and constructing a grating house type image vector data set; 2, identifying walls, doors, windows and scales in the house type image based on a key point detection network, and generating and screening candidate primitives through an axial alignment rule and a constraint condition; step 3, detecting and positioning the endpoints of the scale by using a mode of combining Yolov8 and Shii-Tomasi angular point detection, and performing scale digital character recognition through a pre-trained multi-mode optical character recognition model OFA-OCR (Orthogonal Frequency Amplification-Optical Character Recognition); and step 4, designing and realizing a house type identification and three-dimensional reconstruction system based on a WebGL framework, and supporting rendering of reconstructed wall and door and window views, interaction of a user to the reconstructed house type, and forward and backward functions of house type operation historical records.

The invention discloses a chamfer defect detection method, device and equipment and a storage medium, and relates to the technical field of defect detection.The method comprises the steps that an obtained to-be-detected device image is preprocessed, and chamfer contour points and chamfer contour point pixel coordinates are determined in the preprocessed image; correcting the pixel coordinates of the chamfer contour points, and fitting a chamfer edge straight line by using the corrected pixel coordinates of the chamfer contour points; according to the chamfer edge straight line, screening the corrected chamfer contour point pixel coordinates, and determining a target chamfer end point; and calculating the actual chamfer length based on the target chamfer endpoint, and comparing the actual chamfer length with the preset chamfer length to judge whether the chamfer has defects or not. According to the method, the detection requirements of chamfers of different shapes and sizes can be met by preprocessing and correcting the images, fatigue and subjective judgment errors possibly existing in traditional manual detection are avoided through automatic image processing and data analysis processes, and the detection efficiency is improved.

The invention relates to an X corner detection method based on multilevel matching and feature optimization, and belongs to the technical field of computer vision and image processing. The method comprises the following steps: making a high-precision template corresponding to a predetermined X corner mark, so that the high-precision template can be accurately focused on a corner area; carrying out noise reduction and filtering operations on the image before identifying the mark so as to remove noise interference; carrying out template matching on the preprocessed image by adopting a self-adaptive threshold value and a multi-scale search strategy; and performing multi-level detection on the basis of template matching, including ORB feature extraction, Harris corner optimization and sub-pixel level positioning, and outputting the accurate position of the corner and a feature descriptor. The method can still maintain high precision and stability in a complex scene, is suitable for the fields of visual positioning, three-dimensional reconstruction and the like, and has the advantages of high detection efficiency, strong anti-interference capability, wide applicability and the like.

The invention discloses an aerial view target detection method and system, a storage medium and equipment, and belongs to the field of BEV visual detection map reconstruction.The method comprises the steps that S1, images from different camera visual angles are input, and features of the different camera visual angles are extracted through a backbone network; s2, redesigning a time encoder through deformation and splicing by using long-period time information, and processing the characteristics of different camera visual angles through the designed time encoder; s3, based on a processing result of the step S2, further utilizing dynamic deformable attention to extract spatial features of the image; s4, processing the features obtained in the step S3 by using a feedforward network, and generating a final BEV feature map; and S5, performing target detection by using the final BEV feature map. According to the method, dynamic deformable attention is fused in aerial view angle detection, so that image query is more comprehensive, the calculation overhead is low, meanwhile, time sequence coding is improved, long-time historical information can be better integrated, and the robustness and accuracy of detection are improved.

The invention relates to the field of bottle cap visual inspection, in particular to an AI-based bottle cap 360-degree full-view-angle defect high-speed detection method and system, the method comprises a detection image acquisition step, an image processing and splicing step, a defect identification detection step and a detection result verification step, through multi-sensor collaborative layout, 360-degree dead-corner-free detection of a bottle cap is achieved, and the detection accuracy is improved. And comprehensive detection of the bottle cap is ensured through a high-speed detection mechanism, a clustering model is constructed by utilizing historical defect data, abnormal defects are judged through mahalanobis distance, and the model is dynamically updated based on transfer learning, so that rapid learning of novel defects is realized, and the manual intervention cost is reduced.

The invention discloses a coal cutter roller cutting interference state identification method based on an infrared image, and the method employs an infrared camera to shoot the surrounding environment, effectively prevents the recognition precision from being limited by complex illumination environments, such as underground low illumination, high dust, and strong water mist, and guarantees the stability and definition of a target contour. Then analyzing and processing the infrared image through a deep learning network, determining four corner points of the hydraulic support face guard in the area from the infrared image by adopting a Shii-Tomasi corner point detection algorithm, further accurately calculating the spatial attitude of the hydraulic support face guard, and constructing a three-dimensional trajectory model of the coal mining machine roller; the Kalman filtering is adopted to obtain the predicted poses of the two, and then the cutting interference state is judged through early warning grading, so that the response speed and the safety guarantee capability of the fully-mechanized coal mining system are remarkably improved, the dynamic prediction and real-time response of the cutting interference state are realized, and the efficiency and the precision of cutting interference state recognition are ensured.

The invention discloses a quality detection device for electric lifter processing, and belongs to the technical field of quality detection.The quality detection device comprises a device base, the two sides of the top of the device base are each slidably provided with a sliding frame, the two sliding frames are each in a U shape, and the middle sections of the tops of the two sliding frames are each fixedly provided with a supporting frame; first hydraulic telescopic rods are fixedly arranged on the two sides of the tops of the two supporting frames, first telescopic heads are arranged on the tops of the four first hydraulic telescopic rods in a telescopic mode, and a mounting frame is fixedly arranged between the first telescopic heads arranged on the first hydraulic telescopic rods located on the top of the same supporting frame. And the comprehensiveness and the accuracy of the detection are obviously improved. The system can automatically switch middle, local and corner detection modes by means of a multi-station detection process engine, and realizes 360-degree omnibearing detection of the electric lifter main body in cooperation with the rotation function of the detection placement table and the transverse movement of the sliding frame, thereby avoiding a detection blind area.

The invention discloses a glass edge grinding device and a corner detection method.The glass edge grinding device comprises a first conveying device, the first conveying device is horizontally arranged, a first grinding device, a second grinding device and a corner detection device are sequentially arranged on the first conveying device, and a rotating mechanism is further arranged on the first conveying device; the rotating mechanism is located between the first grinding device and the second grinding device, a guide mechanism is arranged at the tail end of the first conveying device, a crushing device is arranged below the guide mechanism and arranged below the floor, and a second conveying device is further horizontally arranged below the floor. After the two edges of the glass are ground through the first grinding equipment, the glass is turned through the rotating mechanism, so that the other two edges of the glass are ground through the second grinding equipment, the rotating mechanism is simple in structure and convenient to operate, and the working state problem is solved, so that the glass grinding efficiency is improved; unqualified glass is conveyed into the crushing equipment through the guide mechanism to be crushed, and obtained glass residues are conveyed into the smelting furnace through the second conveying equipment to be molten again.

The invention relates to the technical field of ultrasonic medical robots, and discloses a superficial ultrasonic scanning path planning method and device based on an SVD-improved RRT algorithm. The superficial ultrasonic scanning path planning method is applied to ultrasonic medical robot equipment, and specifically comprises the following steps: S101, receiving a data acquisition request sent by a main control end, controlling a Kinect camera to acquire a superficial depth image, processing 20 groups of calibration board images through a Matlab calibration tool box, and executing a corner detection algorithm to identify checkerboard corners, and optimizing and solving the internal reference matrix of the infrared camera. According to the method, SVD plane fitting and an improved RRT algorithm are fused, the precision and efficiency of an ultrasonic scanning path are remarkably improved, in the point cloud processing stage, statistical filtering and voxel gravity center down-sampling synergistically act, point cloud noise is reduced by 90%, meanwhile, more than 95% of three-dimensional features are reserved, a high-fidelity data basis is provided for path planning, and the method has good application prospects. SVD plane fitting forcibly passes through the center of mass, and the error is smaller than or equal to 0.5 mm

The invention provides a centralized multi-robot collaborative SLAM (Simultaneous Localization and Mapping) method based on an FPGA (Field Programmable Gate Array) platform. The method comprises the following steps: acquiring a color image and IMU (Inertial Measurement Unit) data of a surrounding environment by using a motion camera The color image is input into an FPGA platform of the single robot and converted into a grey-scale map, image pyramid scaling, corner detection and feature description are carried out, and the image is transmitted to a processing system of the single robot; the processing system generates observation information and further generates an image frame; generating a key frame and sending the key frame to a central server; the central server side carries out loopback detection on the key frame, carries out map fusion on local maps established by different single robots to obtain a global map, and optimizes the global map; and constructing a three-dimensional dense point cloud map. According to the method, cooperative positioning and three-dimensional mapping of multiple robots can be carried out in a large-scale and complex environment scene, the problem that computing resources and a detectable range of a single robot in the complex scene are limited is solved, and the scale, the precision and the robustness of map construction are improved.

The invention relates to the technical field of image recognition, in particular to a checkerboard corner detection method and system based on image texture, and the method comprises the steps: obtaining a to-be-detected image; converting the image into a grayscale image; projecting a grid in the grayscale image to obtain a grid region; determining a gray jump point on each grid boundary in the grid region; and determining checkerboard angular points according to the gray jump points. According to the technical scheme, the checkerboard angular points are obtained by projecting the grid area and combining judgment of the gray jump points. Compared with the prior art, deep processing such as image separation of the image is avoided, so that the accuracy of the obtained checkerboard angular points is improved.

The invention relates to a calibration block angular point detection method and device based on deep learning and a medium, and the method comprises the steps: collecting a calibration block image which is shot by a vehicle-mounted camera and is used for carrying out camera external parameter calibration, marking points and edges forming a calibration block, and constructing a training set through the marked calibration block image; constructing calibration blocks to form a line segment detection model, and training the model through a training set; inputting a to-be-detected calibration block image into the trained model to obtain a candidate line segment set contained in the to-be-detected calibration block image; selecting four line segments L, L1, L2 and L3 which can enclose a calibration block from the candidate line segment set; and taking four corner points of a quadrangle defined by the four line segments as corner points of the calibration block in the calibration block image to be detected. According to the method, the corner detection precision is improved, and better robustness is achieved.

The invention discloses a linear edge and corner detection method and device and a storage medium, relates to the technical field of machine vision, and is used for accurately acquiring linear edge and corner features. The method comprises the following steps: carrying out convolution on an input image to obtain a gradient magnitude diagram and a gradient direction diagram; acquiring a first image area in the gradient directional diagram; performing first hypothesis testing on the first image region to obtain a first region set and a second region set; dividing the second region set into a first sub-region set and a second sub-region set, and respectively carrying out second hypothesis testing on the first sub-region set and the second sub-region set; if both pass the second hypothesis test, calculating the absolute value of the mean value difference in the gradient direction of the two; determining to add the second region set into the first region set according to the absolute value; performing morphological expansion on the first region set, and calculating the rectangularity of each region; obtaining a linear edge feature region according to a rectangularity clamping control threshold value; and obtaining target linear edge features and angular point features in the linear edge feature region.

The invention discloses a target ground object AI identification method based on a multi-source aerial photo, and relates to the technical field of aerial photo processing, and the method comprises the following steps: S1, collecting a panoramic aerial photo of a to-be-identified region; s2, extracting a first data set by using the gray-level co-occurrence matrix of the panoramic aerial photo; s3, extracting a second data set according to response values of all angular points in the panoramic aerial photo; s4, inputting the first data set and the second data set into a preliminary screening condition model, and taking pixel points meeting the preliminary screening condition model as suspected target ground objects of the to-be-identified region; and S5, processing the suspected target ground feature of the to-be-identified area to obtain an accurate target ground feature. In order to overcome the limitation of a single feature, multiple information sources, such as texture features extracted by a gray-level co-occurrence matrix and structural features extracted by corner detection, are fused, the features of image control points in aerial photography can be described more comprehensively, and the accuracy and robustness of recognition are improved.

The invention relates to the technical field of image processing, in particular to an image data acquisition method and system for a lens plastic part, and the method comprises the steps: obtaining and preprocessing a multi-view surface image, and then evaluating the possibility of a light reflection region according to the gray value and texture features of a pixel point; calculating an angular point response function of each pixel point according to a Harris angular point detection algorithm, presetting a local window size for the pixel points, and correcting the angular point response functions according to the contrast ratio of the pixel points in the local window and the possibility of a reflective area; feature points are obtained according to the corrected angular point response function, multi-view-angle image registration is carried out, the registered multi-view-angle images are fused, a complete image of the surface of the lens plastic part is obtained, and data of the complete image are conveniently stored as a final collection result. According to the method, the problem of analysis difficulty caused by light reflection and feature deficiency is analyzed, so that high-quality acquisition of the full-coverage image on the surface of the lens plastic part is realized.

The invention relates to the technical field of image processing, in particular to an online calibration method and system for hydrological flow data, and solves the technical problem of low flow calculation precision caused by rough window segmentation and illumination interference in the prior art. The method comprises the following steps: collecting a water surface video in a river water flow process, and setting a velocity measurement line area in the water surface video along a water flow direction; according to the gray level distribution of the velocity measurement line area, carrying out illumination state judgment, and according to a judgment result, carrying out image restoration on an illumination abnormal area in the water surface video to obtain a restored water surface video; constructing a space-time image for the repaired water surface video, performing corner detection, clustering corners to divide a plurality of windows, and constructing a window weight for each window; and weighting the texture angle of each window according to the window weight of each window to obtain a weighted texture angle, and calculating hydrological flow data according to the weighted texture angle.

The invention provides an HUD distortion correction method based on cubic Bezier curve fitting. The HUD distortion correction method comprises the following steps: S1, collecting an initial source image and a corresponding initial source distortion image; s2, acquiring control point coordinates of the initial source image and the initial source distortion image by using a corner detection algorithm; s3, according to the control point coordinates of the initial source image and the initial source distorted image, obtaining a predistortion function of the initial source image by using cubic Bezier curve fitting; s4, performing predistortion on the initial source image by using the predistortion function to obtain a predistortion image; and S5, replacing the initial source image with the pre-distorted image, and inputting the pre-distorted image into the HUD device to obtain a corrected HUD image. According to the method, the predistortion image is projected to the automobile front windshield, so that the relative coordinate points of the HUD image imaged on the automobile front windshield and the initial source image coincide, the display effect of the improved image is consistent with the display effect of the initial source image, and the driving feeling of a driver is greatly improved.

The invention provides a visual detection method and system, a storage medium and electronic equipment. The visual detection method comprises the following steps: acquiring images shot by a plurality of vehicle-mounted cameras, and acquiring a panoramic look-around image according to the images; the image and the panoramic look-around image are input into a detection model, a detection result is obtained through the detection model, and the detection model comprises a parking space corner detection model, a drivable area detection model and a target object detection model; the detection result comprises a parking space corner detection result, a driving area detection result and a target object detection result; packaging the detection result, CAN signal data of the vehicle body and other to-be-transmitted data into a standardized data format to obtain a packaged file; and transmitting the packaged file to an upper computer. The visual detection method has relatively high recognition accuracy.

The invention provides a multi-unmanned aerial vehicle cooperative path planning method for mountain monitoring, and the method comprises the steps: collecting an image of a to-be-detected mountain through a set unmanned aerial vehicle cluster formation, carrying out the feature point extraction of the to-be-detected mountain through a Harris corner detection algorithm, marking the feature points of a mountain top, the feature points of a mountain depression and the feature points of a mountain foot, and carrying out the positioning of the to-be-detected mountain. The method comprises the steps that feature points of a to-be-detected mountain are obtained, coordinates of the feature points in a world coordinate system are obtained through calculation, then a three-dimensional model of the to-be-detected mountain is obtained, then the feature points serve as task detection points to be distributed to an unmanned aerial vehicle cluster through a greedy algorithm, and three-dimensional path planning is conducted on all unmanned aerial vehicles distributed with detection tasks according to an A-Star algorithm through the three-dimensional model of the to-be-detected mountain. According to the invention, the unmanned aerial vehicles are used for automatic formation and cooperative acquisition of the to-be-detected mountain image, the labor investment is reduced, the system can adapt to different terrains and environment conditions, the data acquisition coverage range is wider, the working efficiency is improved, and powerful support is provided for monitoring of natural disasters in mountainous areas.

The invention discloses a robot multi-component low-cost high-precision calibration method based on a consumer-level camera, and the method comprises the steps: photographing a calibration plate through the consumer-level camera, and optimizing the internal parameters of the camera through corner detection and nonlinear optimization; a camera is installed at the tail end of the mechanical arm, about 50 sets of calibration plate images are shot, the pose is obtained by means of forward kinematics of the mechanical arm, and the pose relation between the camera and the mechanical arm is analyzed through the formula AX = XB; then the chassis and the camera on the mechanical arm are used for joint calibration, the calibration relation between the mechanical arm and the trolley is solved by moving the trolley and the mechanical arm observation calibration plate, and finally the calibration precision is verified by calculating the re-projection error. The method does not need to depend on high-precision equipment and a professional calibration field, calibration can be completed by using a common consumer camera and a simple calibration board, the method can adapt to various cameras, the requirement for computing equipment is not high, and the requirement can be met by a common workstation. The method can be carried out in a dynamic scene, the calibration time is within 1 hour, and the precision is within 5 mm.

The invention provides a dense point cloud-based building feature line extraction method, and belongs to the technical field of building feature line extraction, and the method comprises the following steps: carrying out the data preprocessing of point cloud data, carrying out the edge detection and corner detection of the preprocessed point cloud data, and carrying out the linear segment detection and curve segment detection of feature detection data, and performing line segment connection and removal processing on the line segment detection data, establishing a mapping relation table of optimized data, performing three-dimensional data model reconstruction on the building, and generating a three-dimensional model of the building. The cloud point data is added possibly at the beginning, errors and defects caused by environmental influence can be avoided, raindrop data are richer and more perfect, then the cloud point data are denoised, summary recognition of an initial model is achieved, irrelevant points are removed in the later period, follow-up data processing is more accurate, and the accuracy of raindrop data processing is improved. And through edge and curve identification, the boundary of the model can be better planned, and the subsequent processing efficiency is higher.

The invention relates to the field of industrial detection, and discloses a seamless steel tube surface defect detection device, which comprises a support frame, which is used as an equipment foundation chassis and is used for supporting the stable operation of the whole equipment; the electric guide rail is arranged at the top of the supporting frame and used for driving the steel pipe to operate through a stepping motor; the clamping mechanism is arranged at the top of the electric guide rail and used for being driven by a stepping motor of the electric guide rail to drive the steel pipe to move; and the annular guide rail is arranged at the top of the support frame and is used for assisting the high-definition camera to flexibly change the photographing angle. Through a multi-angle and multi-direction full-coverage acquisition mode of a high-definition camera, an annular guide rail and an adjustable LED light source module, dead-corner-free detection on the surface of the steel pipe is realized, and meanwhile, defects such as cracks, recesses and scratches are accurately identified by combining image preprocessing technologies such as Gaussian filtering and histogram equalization and a deep learning algorithm; the detection precision is greatly improved, and missing detection and false detection are effectively avoided.

The invention provides a corner detection method and device, computer equipment and a program product, and the method comprises the steps: detecting each corner in a checkerboard image through employing a pre-trained corner detection network, and obtaining a first detection frame corresponding to each corner; performing image optimization operation on the detection frame images corresponding to the first detection frames to obtain optimized images; according to the gradient of each pixel point in the optimized image, screening out candidate saddle points conforming to saddle point features from each pixel point; and according to the first coordinate of each candidate saddle point in the optimized image, screening out a target corner point in the first detection frame from each candidate saddle point.

The invention relates to the technical field of image processing, in particular to a shaft part forging forming monitoring method and system. The method comprises the following steps: acquiring characteristic values of pixel points in a surface image of a formed shaft part in a plurality of channels in an HSV color space, and gradient direction angles of the pixel points; determining the local consistency of the pixel points; determining a weighted Euclidean distance between the pixel points; clustering all pixel points according to the weighted Euclidean distance to obtain a plurality of clusters; determining the gradient saliency and the number of candidate angular points of the image region where the clustering cluster is located; and on the basis of the number of the candidate corner points, the candidate corner points are obtained through a Harris corner point detection algorithm, and whether forging of the shaft part is qualified or not is measured according to the Euclidean distance between the candidate corner points. According to the method, the problem of inaccurate recognition caused by large feature difference of different geometric regions is solved through the number of the candidate angular points of the image region where the adaptive clustering cluster is located, and the accuracy of monitoring the forging quality of the shaft parts is improved.