CN117589065A - Detection method for size of interface of special-shaped shaft - Google Patents
Detection method for size of interface of special-shaped shaft Download PDFInfo
- Publication number
- CN117589065A CN117589065A CN202311545370.9A CN202311545370A CN117589065A CN 117589065 A CN117589065 A CN 117589065A CN 202311545370 A CN202311545370 A CN 202311545370A CN 117589065 A CN117589065 A CN 117589065A
- Authority
- CN
- China
- Prior art keywords
- model
- algorithm
- training
- size
- yolov8
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 41
- 238000000034 method Methods 0.000 claims abstract description 24
- 238000012549 training Methods 0.000 claims description 36
- 238000012545 processing Methods 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 8
- 238000006243 chemical reaction Methods 0.000 claims description 7
- 238000004519 manufacturing process Methods 0.000 claims description 7
- 238000012937 correction Methods 0.000 claims description 6
- 230000006870 function Effects 0.000 claims description 6
- 238000002372 labelling Methods 0.000 claims description 6
- 238000012795 verification Methods 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 4
- 238000000137 annealing Methods 0.000 claims description 3
- 230000002146 bilateral effect Effects 0.000 claims description 3
- 230000000694 effects Effects 0.000 claims description 3
- 238000002474 experimental method Methods 0.000 claims description 3
- 230000008014 freezing Effects 0.000 claims description 3
- 238000007710 freezing Methods 0.000 claims description 3
- 238000009499 grossing Methods 0.000 claims description 3
- 230000003190 augmentative effect Effects 0.000 claims description 2
- 238000005259 measurement Methods 0.000 abstract description 16
- 230000001788 irregular Effects 0.000 abstract description 8
- 238000012360 testing method Methods 0.000 description 5
- 238000005299 abrasion Methods 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000000691 measurement method Methods 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 230000001627 detrimental effect Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000011897 real-time detection Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000004513 sizing Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of target detection methods, in particular to a detection method for the size of an interface of an irregular shaft. The method utilizes a YOLOv8 target detection algorithm to identify two parts of a smooth straight line and an adjacent inclined straight line of a part in a picture, namely a target frame 1 and a target frame 2, and then carries out scale and angle estimation on the obtained outline length width height dimension of the target frame of a designated part of the part, wherein the scale and angle estimation mainly comprises the steps of realizing geometric calibration by estimating absolute dimensions, and recovering information such as the orientation (horizontal plane in an image), the view field, the absolute height of the camera and the ground and the like of the camera. Finally, with this information, two-dimensional measurements in the image can be converted into three-dimensional measurements in real space. Fitting the contour length and angle by using the gradient descent algorithm SGDRegresor and the truth value through the mathematical relationship between the two types of target boxes and the standard value.
Description
Technical Field
The invention relates to the technical field of target detection methods, in particular to a detection method for the size of an interface of an irregular shaft.
Background
Visual sizing is a common need in industrial inspection. However, monocular vision measurements have low accuracy and large field of view (FOV), while multi-vision measurements involve multi-source information fusion and incur a significant computational overhead. Therefore, we propose a planar dimension measurement optimization method that measures a height difference that can reduce the error. Conventional measurements typically require manual operation using calipers, screw micrometers, and other tools. However, with advances in computer and image processing technology, visual measurement has become one of the most promising measurement methods in engineering applications. The main categories are two: a multi-vision measurement method and a monocular vision measurement method. Multi-vision measurement uses triangulation to obtain planar coordinates of image feature points from several different views. It is used in many fields due to its large field of view (FOV) and high accuracy. For example, genovese et al propose methods that utilize planar mirrors and phase targets, but the calibration process is overly cumbersome and detrimental to engineering applications.
Multi-vision methods involve the processing and correlation of data from multiple cameras, requiring camera time synchronization and spatial calibration to ensure consistency of visual information. Parallax computation is a key step in a multi-vision method, involves pixel-level matching and searching, and has high computational complexity. Although the measurement accuracy of the multi-vision method is high, it requires a cumbersome calibration process at the initial stage of measurement and a recalibration whenever the camera position is changed. This aspect makes the monocular approach advantageous in terms of convenience and flexibility.
In monocular measurement, the conversion parameters from the world coordinate system to the image coordinate system are derived by simple calibration. The monocular measurement has the characteristics of simple structure and high efficiency, and is suitable for industrial equipment with limited resources. In the prior art, the size and angle of the special-shaped part are detected, and the direct measurement is generally carried out by using a spot check, or the whole irregular part is converted into a regular rectangle, and the length and width of the part are estimated approximately by measuring the length and width of the rectangle. However, the dimension and angle of a certain part of the special-shaped part cannot be detected, and the two methods are suitable for detection in a laboratory and are not beneficial to the rapid size detection in a complex environment on a production line.
Disclosure of Invention
The invention aims to provide a detection method for the size of an interface of an irregular shaft, so as to improve the robustness of detection means and part detection, and a complicated camera calibration process is not needed, so that the method is more beneficial to engineering application; the calibration process is to estimate the information such as the orientation, focal length, object distance and the like of the camera according to the original YOLOv 8; the input picture does not need to be subjected to camera rectification.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
the method utilizes a YOLOv8 target detection algorithm to identify two parts of a smooth straight line and an adjacent inclined straight line of a part in a picture, namely a target frame 1 and a target frame 2, and then carries out scale and angle estimation on the obtained outline length width height dimension of the target frame of a designated part of the part, wherein the scale and angle estimation mainly comprises the steps of realizing geometric calibration by estimating absolute dimensions, and recovering information such as the orientation (horizontal plane in an image), the view field, the absolute height of the camera and the ground and the like of the camera. Finally, with this information, two-dimensional measurements in the image can be converted into three-dimensional measurements in real space. The fitting contour length and angle are carried out by using the gradient descent algorithm SGDRegresor and true value through the mathematical relationship between the two types of target frames and the standard value, and the specific contents are as follows.
A detection method for the size of an interface of an irregular shaft comprises the following steps:
s1, acquiring an image of a special-shaped part on a laboratory bench by using a CMOS camera for manufacturing a data set, and labeling the data set by using labeling software; augmenting the data set, and dividing the data set into a training set and a verification set;
s2, processing the data set obtained in the S1, and inputting the processed data set into the YOLOv8 for training to obtain a model v8.Pt file;
s3, detecting pixel sizes and rotation angles of the target frames 1 and 2 by using a camera parameter estimation model;
s4, selecting a gradient descent algorithm SGDRegressor as an error compensation model, and using the error compensation model to approximate the angle and the size of a target frame in the two-dimensional picture to the size of a part to be detected, so as to indirectly represent the angle of the measured part of the two-dimensional picture; estimating the angle and the size of the two-dimensional space and the error of the size and the angle of the three-dimensional space, and training an error compensation model to obtain a regression_model.
S4, adding a conversion and correction module into the YOLOv8, constructing a camera parameter estimation-error compensation detection model based on a YOLOv 8-SGDRegresor algorithm by combining a camera parameter estimation model and an error compensation model, training the model, and then inputting a model v8.Pt file and a regression_model. Pt file to perform size conversion and correction on the predicted value to obtain a real size and an angle.
Preferably, the S1 specifically includes the following:
the image of the special-shaped part on the experiment table is acquired by using the CMOS camera and is used for manufacturing an image dataset of the special-shaped shaft interface part, the obtained dataset is amplified, and the image dataset is amplified according to 8:2, dividing the data set into a training section and a verification set, and marking the adjacent two parts of the detection part in all the images with region and category information.
Preferably, the S2 specifically includes the following:
s2.1, determining a bilateral threshold according to the histogram features of the gray level image;
s2.2, respectively determining threshold values for the intervals of the pixel value of the part smaller than 20 and the pixel value of the part larger than 240: taking average of the maximum pixel value and the minimum pixel value in the interval of the pixel value of the part smaller than 20 as a minimum threshold value; for the section with the pixel value of the part being greater than 240, calculating a threshold value with the number of pixels with the pixel value being greater than 240 being a median as a maximum threshold value;
s2.3, performing Gaussian filtering processing on a section between the maximum threshold value and the minimum threshold value in the S2.2;
s2.4, carrying out edge extraction on the image by utilizing a Sobel operator;
and S2.5, labeling the picture after the edge extraction in the S2.4, and training in Yolov8 to obtain a model v8.Pt file.
Preferably, the training of the model for camera parameter estimation-error compensation detection based on the YOLOv 8-sgdregsor algorithm in S4 specifically includes the following: setting training parameters, wherein the training iteration times are 500 times, training is performed in a freezing mode for the first 50 times, and the learning rate is 0.001; the learning rate of the last 250 iterations is 0.0001; the training process adopts an annealing cosine algorithm and a label smoothing algorithm.
Preferably, the yolov8n_cbam.yaml camera parameter estimation-error compensation detection model based on the yolov8n_sms_object algorithm in S4 includes a kalman filter module, and the kalman filter module is provided with yolov8n_smalljobject.yaml of a CBAM attention mechanism, so as to enhance the detection effect on small targets.
Preferably, the model for camera parameter estimation-error compensation detection based on YOLOv 8-sgdregsor algorithm in S4 further includes a loss function module, where the loss function module is specifically classified into loss: lcls_loss and bounding box regression loss: box_loss+ dfl _loss.
Compared with the prior art, the invention provides a Raw domain image and video mole pattern removing method based on channel modulation and spatial modulation, which has the following beneficial effects:
(1) The invention can rapidly and accurately detect the size and the angle of the part in the picture;
(2) By utilizing the part size and angle detection method provided by the invention, the size and angle of a certain part of the part can be estimated by directly inputting an uncorrected picture without calibrating and correcting a camera;
(3) Compared with the traditional method for detecting the size of the part, the invention is more suitable for detecting the size and angle of the part on an industrial assembly line.
Drawings
FIG. 1 is a histogram of a gray scale image of a part as referred to in example 1 of the present invention;
FIG. 2 is a histogram of pixel values of less than 20 for the part of example 1 of the present invention;
FIG. 3 is a histogram of pixel values of the part of example 1 above 240;
FIG. 4 is a flow chart of the measurement and angle detection of the profiled element in example 2 of the present invention;
fig. 5 is a drawing showing the dimensions and angle detection recognition of the parts mentioned in embodiment 2 of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments.
The research is based on the current most advanced real-time monitoring algorithm YOLOv8 target detection algorithm, is combined with an improved EfficientNet convolutional neural network, and improves the combined network. The pretreatment operation is carried out on the obtained conveyor belt abrasion picture, the pretreatment technology adopts a Retinex image enhancement algorithm to carry out brightness enhancement, detail protection, color protection, naturalness recovery, image denoising, detail extraction and the like on the image, so that the definition of the belt abrasion picture is improved. And finally, deploying the improved algorithm to a real-time detection device of the conveyor to detect the belt of the conveyor, and accurately obtaining the abrasion part of the conveyor belt and the abrasion degree of each part. The method for detecting the size of the interface of the irregular shaft, which is proposed by the invention, is described below with reference to specific examples.
Example 1:
the invention provides a method for detecting the size of an interface of an irregular shaft, which comprises the following steps:
s1, acquiring an image of a special-shaped part on a laboratory bench by using a CMOS camera for manufacturing a data set, and labeling the data set by using labeling software; the data set is amplified and divided into a training set and a verification set, and the method specifically comprises the following steps:
the image of the special-shaped part on the experiment table is acquired by using the CMOS camera and is used for manufacturing an image dataset of the special-shaped shaft interface part, the obtained dataset is amplified, and the image dataset is amplified according to 8:2 dividing the data set into a training section and a verification set according to the proportion, and marking the adjacent two parts of the detection part in all the images with region and category information;
s2, processing the data set obtained in the S1, and inputting the processed data set into the Yolov8 for training to obtain a model v8.Pt file, wherein the method specifically comprises the following steps:
s2.1, referring to FIG. 1, determining a bilateral threshold according to histogram features of the gray image;
s2.2, respectively determining threshold values for the intervals of the pixel value of the part smaller than 20 and the pixel value of the part larger than 240: referring to fig. 2-3, taking an average of the maximum pixel value and the minimum pixel value as a minimum threshold value in a section where the pixel value of the component is less than 20; for the section with the pixel value of the part being greater than 240, calculating a threshold value with the number of pixels with the pixel value being greater than 240 being a median as a maximum threshold value;
s2.3, performing Gaussian filtering processing on a section between the maximum threshold value and the minimum threshold value in the S2.2;
s2.4, carrying out edge extraction on the image by utilizing a Sobel operator;
s2.5, labeling the picture after the edge extraction in the S2.4, and training in YOLOv8 to obtain a model v8.Pt file;
s3, detecting pixel sizes and rotation angles of the target frames 1 and 2 by using a camera parameter estimation model;
s4, selecting a gradient descent algorithm SGDRegressor as an error compensation model, and using the error compensation model to approximate the angle and the size of a target frame in the two-dimensional picture to the size of a part to be detected, so as to indirectly represent the angle of the measured part of the two-dimensional picture; estimating the angle and the size of the two-dimensional space and the error of the size and the angle of the three-dimensional space, and training an error compensation model to obtain a regression_model.
S4, adding a conversion and correction module into the YOLOv8, constructing a camera parameter estimation-error compensation detection model based on a YOLOv 8-SGDRegresor algorithm by combining a camera parameter estimation model and an error compensation model, training the model, and then inputting a model v8.Pt file and a regression_model. Pt file to perform size conversion and correction on the predicted value to obtain a real size and an angle;
the model training specifically comprises the following steps: setting training parameters, wherein the training iteration times are 500 times, training is performed in a freezing mode for the first 50 times, and the learning rate is 0.001; the learning rate of the last 250 iterations is 0.0001; the training process adopts an annealing cosine algorithm and a label smoothing algorithm;
the camera parameter estimation-error compensation detection model based on the YOLOv8-SGDRegressor algorithm further comprises a Kalman filter module and a loss function module, wherein the Kalman filter module is provided with yolov8n_cbam.yaml of a CBAM attention mechanism, and simultaneously is additionally provided with yolov8n_small_object.yaml to enhance the detection effect on small targets; the loss function module is specifically classified loss: lcls_loss and bounding box regression loss: box_loss+ dfl _loss.
Based on example 1 but with the difference that,
referring to fig. 4, the present embodiment provides a method for detecting the size of an interface of an irregular shaft, which includes the following steps:
step 1: preparing at least 800 photos as a dataset;
step 2: data preprocessing: preprocessing a training and testing data set, and carrying out graying, gaussian filtering, binarization and sobel edge extraction processing on the data set; selecting proper part pictures, marking the size and angle to be detected, and according to 8:2 as training and testing data sets;
step 3: building a model and training: utilizing YOLOv8 to predict the length, width, height and angle of a part to be measured of the special-shaped part in the picture by using the 2D image to obtain coordinates of an object, and finishing detection of the target frame 1 and the target frame 2, namely identifying the target frame 1 and the target frame 2;
step 4: carrying out regression prediction on the predicted value and the true value by a random gradient descent algorithm SGDRegressor on the size and the angle of the part to be measured; the results are shown in FIG. 5:
obtain the length l of the target frame 1 1 Sum width w 1 Length l of target frame 2 2 . The frame definition of YOLOv8 is [ x_c, y_c, longside, shortside,]the modified angle information is stored in the computer, and can be used for storing [ x_c, y_c, longside, shortside]Regarding the angle as a horizontal target frame, outputting the angle of the target frame 1 and the angle of the target frame 2, and obtaining the angle of the special-shaped part;
step 5: model test: and testing the trained Yolov8+SGDRegressor model by using a test data set, and detecting the size and angle of the part to obtain a detection result.
The foregoing is only a preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art, who is within the scope of the present invention, should make equivalent substitutions or modifications according to the technical scheme of the present invention and the inventive concept thereof, and should be covered by the scope of the present invention.
Claims (6)
1. The method for detecting the size of the interface of the special-shaped shaft is characterized by comprising the following steps:
s1, acquiring an image of a special-shaped part on a laboratory bench by using a CMOS camera for manufacturing a data set, and labeling the data set by using labeling software; augmenting the data set, and dividing the data set into a training set and a verification set;
s2, processing the data set obtained in the S1, and inputting the processed data set into the YOLOv8 for training to obtain a model v8.Pt file;
s3, detecting pixel sizes and rotation angles of the target frames 1 and 2 by using a camera parameter estimation model;
s4, selecting a gradient descent algorithm SGDRegressor as an error compensation model, and using the error compensation model to approximate the angle and the size of a target frame in the two-dimensional picture to the size of a part to be detected, so as to indirectly represent the angle of the measured part of the two-dimensional picture; estimating the angle and the size of the two-dimensional space and the error of the size and the angle of the three-dimensional space, and training an error compensation model to obtain a regression_model.
S4, adding a conversion and correction module into the YOLOv8, constructing a camera parameter estimation-error compensation detection model based on a YOLOv 8-SGDRegresor algorithm by combining a camera parameter estimation model and an error compensation model, training the model, and then inputting a model v8.Pt file and a regression_model. Pt file to perform size conversion and correction on the predicted value to obtain a real size and an angle.
2. The belt wear state detection method based on the improved YOLOv8-EfficientNet algorithm of claim 1, wherein S1 specifically comprises the following:
the image of the special-shaped part on the experiment table is acquired by using the CMOS camera and is used for manufacturing an image dataset of the special-shaped shaft interface part, the obtained dataset is amplified, and the image dataset is amplified according to 8:2, dividing the data set into a training section and a verification set, and marking the adjacent two parts of the detection part in all the images with region and category information.
3. The belt wear state detection method based on the improved YOLOv8-EfficientNet algorithm of claim 2, wherein S2 specifically comprises the following:
s2.1, determining a bilateral threshold according to the histogram features of the gray level image;
s2.2, respectively determining threshold values for the intervals of the pixel value of the part smaller than 20 and the pixel value of the part larger than 240: taking average of the maximum pixel value and the minimum pixel value in the interval of the pixel value of the part smaller than 20 as a minimum threshold value; for the section with the pixel value of the part being greater than 240, calculating a threshold value with the number of pixels with the pixel value being greater than 240 being a median as a maximum threshold value;
s2.3, performing Gaussian filtering processing on a section between the maximum threshold value and the minimum threshold value in the S2.2;
s2.4, carrying out edge extraction on the image by utilizing a Sobel operator;
and S2.5, labeling the picture after the edge extraction in the S2.4, and training in Yolov8 to obtain a model v8.Pt file.
4. The belt wear state detection method based on the improved YOLOv8-EfficientNet algorithm of claim 1, wherein the training of the camera parameter estimation-error compensation detection model based on the YOLOv 8-sgdregsor algorithm in S4 specifically comprises the following: setting training parameters, wherein the training iteration times are 500 times, training is performed in a freezing mode for the first 50 times, and the learning rate is 0.001; the learning rate of the last 250 iterations is 0.0001; the training process adopts an annealing cosine algorithm and a label smoothing algorithm.
5. The belt wear state detection method based on the improved YOLOv8-EfficientNet algorithm according to claim 1, wherein a kalman filter module is included in the camera parameter estimation-error compensation detection model based on the YOLOv 8-sgdregsor algorithm in S4, and the kalman filter module is provided with yolov8n_cbam. Yaml of a CBAM attention mechanism, and is additionally provided with yolov8n_small_object. Yaml to enhance the detection effect on small objects.
6. The belt wear state detection method based on the improved YOLOv8-EfficientNet algorithm of claim 5, wherein the camera parameter estimation-error compensation detection model based on the YOLOv 8-sgdregesensor algorithm in S4 further comprises a loss function module, and the loss function module is specifically classified into loss: lcls_loss and bounding box regression loss: box_loss+ dfl _loss.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311545370.9A CN117589065B (en) | 2023-11-20 | 2023-11-20 | Detection method for size of interface of special-shaped shaft |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311545370.9A CN117589065B (en) | 2023-11-20 | 2023-11-20 | Detection method for size of interface of special-shaped shaft |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117589065A true CN117589065A (en) | 2024-02-23 |
| CN117589065B CN117589065B (en) | 2024-09-20 |
Family
ID=89921198
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311545370.9A Active CN117589065B (en) | 2023-11-20 | 2023-11-20 | Detection method for size of interface of special-shaped shaft |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117589065B (en) |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106095729A (en) * | 2016-06-16 | 2016-11-09 | 广东工业大学 | Optical grating ruler measurement error dynamic compensation method based on degree of depth study |
| CN110689137A (en) * | 2019-09-24 | 2020-01-14 | 网易传媒科技（北京）有限公司 | Parameter determination method, system, medium, and electronic device |
| CN111563415A (en) * | 2020-04-08 | 2020-08-21 | 华南理工大学 | Binocular vision-based three-dimensional target detection system and method |
| WO2021244079A1 (en) * | 2020-06-02 | 2021-12-09 | 苏州科技大学 | Method for detecting image target in smart home environment |
| CN114119553A (en) * | 2021-11-28 | 2022-03-01 | 长春理工大学 | Binocular vision out-of-plane circular hole detection method based on cross laser |
| CN114445706A (en) * | 2022-01-25 | 2022-05-06 | 南京工程学院 | Power transmission line target detection and identification method based on feature fusion |
| CN114549629A (en) * | 2022-02-23 | 2022-05-27 | 中国海洋大学 | Method for estimating three-dimensional pose of target by underwater monocular vision |
| CN115375672A (en) * | 2022-09-14 | 2022-11-22 | 河北工程大学 | Coal gangue detection method based on improved YOLOv4 algorithm |
| CN115565024A (en) * | 2022-09-09 | 2023-01-03 | 中铁十九局集团矿业投资有限公司 | Ore lumpiness detection method based on YOLOV5 image recognition algorithm and depth camera |
| CN115953446A (en) * | 2022-12-15 | 2023-04-11 | 际络科技(上海)有限公司 | Depth estimation method and device based on error correction and electronic equipment |
| CN116189140A (en) * | 2022-12-20 | 2023-05-30 | 重庆邮电大学 | Binocular vision-based vehicle three-dimensional target detection algorithm |
| CN116389913A (en) * | 2023-04-07 | 2023-07-04 | 武汉纺织大学 | Image exposure correction method based on depth curve estimation |
| CN116659396A (en) * | 2023-06-19 | 2023-08-29 | 南昌航空大学 | Helicopter airborne rotor blade tip displacement measurement method based on point light source detection |
-
2023
- 2023-11-20 CN CN202311545370.9A patent/CN117589065B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106095729A (en) * | 2016-06-16 | 2016-11-09 | 广东工业大学 | Optical grating ruler measurement error dynamic compensation method based on degree of depth study |
| CN110689137A (en) * | 2019-09-24 | 2020-01-14 | 网易传媒科技（北京）有限公司 | Parameter determination method, system, medium, and electronic device |
| CN111563415A (en) * | 2020-04-08 | 2020-08-21 | 华南理工大学 | Binocular vision-based three-dimensional target detection system and method |
| WO2021244079A1 (en) * | 2020-06-02 | 2021-12-09 | 苏州科技大学 | Method for detecting image target in smart home environment |
| CN114119553A (en) * | 2021-11-28 | 2022-03-01 | 长春理工大学 | Binocular vision out-of-plane circular hole detection method based on cross laser |
| CN114445706A (en) * | 2022-01-25 | 2022-05-06 | 南京工程学院 | Power transmission line target detection and identification method based on feature fusion |
| CN114549629A (en) * | 2022-02-23 | 2022-05-27 | 中国海洋大学 | Method for estimating three-dimensional pose of target by underwater monocular vision |
| CN115565024A (en) * | 2022-09-09 | 2023-01-03 | 中铁十九局集团矿业投资有限公司 | Ore lumpiness detection method based on YOLOV5 image recognition algorithm and depth camera |
| CN115375672A (en) * | 2022-09-14 | 2022-11-22 | 河北工程大学 | Coal gangue detection method based on improved YOLOv4 algorithm |
| CN115953446A (en) * | 2022-12-15 | 2023-04-11 | 际络科技(上海)有限公司 | Depth estimation method and device based on error correction and electronic equipment |
| CN116189140A (en) * | 2022-12-20 | 2023-05-30 | 重庆邮电大学 | Binocular vision-based vehicle three-dimensional target detection algorithm |
| CN116389913A (en) * | 2023-04-07 | 2023-07-04 | 武汉纺织大学 | Image exposure correction method based on depth curve estimation |
| CN116659396A (en) * | 2023-06-19 | 2023-08-29 | 南昌航空大学 | Helicopter airborne rotor blade tip displacement measurement method based on point light source detection |
Non-Patent Citations (2)
| Title |
|---|
| 姜卫: "基于改进型YOLOv3和模型轻量化的锈蚀检测及应用", 《中国优秀硕士学位论文全文库▪工程科技Ⅰ辑》, 31 March 2022 (2022-03-31), pages 1 - 68 * |
| 杨立洁等: "基于EfficientNet-YOLOv3的多氧压表识别方法", 《制造业自动化》, vol. 45, no. 8, 31 August 2023 (2023-08-31), pages 43 - 48 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117589065B (en) | 2024-09-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Fan et al. | Pothole detection based on disparity transformation and road surface modeling | |
| CN104408460B (en) | A kind of lane detection and tracking detection method | |
| CN107014294B (en) | Contact net geometric parameter detection method and system based on infrared image | |
| US11651581B2 (en) | System and method for correspondence map determination | |
| CN108470356B (en) | Target object rapid ranging method based on binocular vision | |
| JP6899189B2 (en) | Systems and methods for efficiently scoring probes in images with a vision system | |
| CN104075659A (en) | Three-dimensional imaging recognition method based on RGB structure light source | |
| CN105894521A (en) | Sub-pixel edge detection method based on Gaussian fitting | |
| CN110378924A (en) | Level set image segmentation method based on local entropy | |
| CN111160477B (en) | Image template matching method based on feature point detection | |
| CN115096206A (en) | Part size high-precision measurement method based on machine vision | |
| CN109671084B (en) | Method for measuring shape of workpiece | |
| CN103337080A (en) | Registration technology of infrared image and visible image based on Hausdorff distance in gradient direction | |
| CN116977341B (en) | Dimension measurement method and related device | |
| CN117589065B (en) | Detection method for size of interface of special-shaped shaft | |
| CN117078608B (en) | Double-mask guide-based high-reflection leather surface defect detection method | |
| CN115184362B (en) | Rapid defect detection method based on structured light projection | |
| CN116188826A (en) | Template matching method and device under complex illumination condition | |
| Belhaoua et al. | Estimation of 3D reconstruction errors in a stereo-vision system | |
| CN108280815B (en) | Geometric correction method for monitoring scene structure | |
| Liu et al. | Outdoor camera calibration method for a GPS & camera based surveillance system | |
| Wang et al. | Computer vision methods for automating high temperature steel section sizing in thermal images | |
| Rao et al. | A real-time auto-recognition method for pointer-meter under uneven illumination | |
| Agarwal et al. | Image Processing Based Technique to Detect Contour Area | |
| Bhardwaj | Detection and verification of missing components in SMD using AOI techniques |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Country or region after: China Address after: 056009 No. 19, Taiji Road, Handan economic and Technological Development Zone, Hebei Province Applicant after: HEBEI University OF ENGINEERING Address before: No. 199, Guangming South Street, Handan, Hebei 056009 Applicant before: HEBEI University OF ENGINEERING Country or region before: China |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |