CN117197743A - Belt longitudinal tearing detection method based on multi-frame two-dimensional point cloud identification - Google Patents
Belt longitudinal tearing detection method based on multi-frame two-dimensional point cloud identification Download PDFInfo
- Publication number
- CN117197743A CN117197743A CN202311180502.2A CN202311180502A CN117197743A CN 117197743 A CN117197743 A CN 117197743A CN 202311180502 A CN202311180502 A CN 202311180502A CN 117197743 A CN117197743 A CN 117197743A
- Authority
- CN
- China
- Prior art keywords
- belt
- point cloud
- neural network
- frame
- dimensional point
- 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.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 40
- 238000012549 training Methods 0.000 claims abstract description 30
- 238000003062 neural network model Methods 0.000 claims abstract description 20
- 238000013528 artificial neural network Methods 0.000 claims abstract description 11
- 238000004364 calculation method Methods 0.000 claims abstract description 7
- 238000007781 pre-processing Methods 0.000 claims abstract description 7
- 238000000034 method Methods 0.000 claims description 26
- 238000001914 filtration Methods 0.000 claims description 12
- 238000013136 deep learning model Methods 0.000 claims description 7
- 238000012360 testing method Methods 0.000 claims description 6
- 238000012795 verification Methods 0.000 claims description 6
- 230000005856 abnormality Effects 0.000 claims description 4
- 230000002159 abnormal effect Effects 0.000 claims description 3
- 230000006870 function Effects 0.000 claims description 3
- 230000001537 neural effect Effects 0.000 claims description 3
- 230000000717 retained effect Effects 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 230000000007 visual effect Effects 0.000 claims description 3
- 230000009286 beneficial effect Effects 0.000 abstract description 5
- 239000000463 material Substances 0.000 description 5
- 238000013135 deep learning Methods 0.000 description 3
- 239000000428 dust Substances 0.000 description 2
- 239000004677 Nylon Substances 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 238000012806 monitoring device Methods 0.000 description 1
- 229920001778 nylon Polymers 0.000 description 1
- 230000000149 penetrating effect Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000011179 visual inspection Methods 0.000 description 1
Landscapes
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of computer vision detection, and discloses a belt longitudinal tearing detection method based on multi-frame two-dimensional point cloud identification, which comprises neural network model training and neural network model prediction; the neural network model training comprises the following steps: s1: preparing data required for training; s2: building a neural network to be trained; s3: preprocessing the data in the step S1; s4: training a neural network model; s5: training iteration and loss calculation; s6: obtaining a network model weight; the neural network model prediction comprises the following steps: k1: two-dimensional point cloud; k2: reading a point cloud frame; s3: and (5) running neural network prediction. The laser radar has high precision, the laser radar has high ranging precision, can accurately detect the surface change and the irregularity of the belt, and is beneficial to timely finding out the tearing phenomenon of the belt.
Description
Technical Field
The invention relates to the technical field of computer vision detection, in particular to a belt longitudinal tearing detection method based on multi-frame two-dimensional point cloud identification.
Background
In the field of development of belt longitudinal tear monitoring devices, a number of different solutions have been proposed both at home and abroad. Existing belt tear detection devices can be largely divided into two main categories: one is to use mechanical device to detect, including linear detector and leakage detector; the other type is detection by adopting a computer vision technology, and mainly comprises a computer vision detection method based on deep learning and a vision detection method based on line laser and a CCD camera.
1. Line detector
The detector is arranged below the groove-shaped belt, a metal wire or nylon wire is pulled along the outline of the belt, and a spring type limit switch is arranged at one end of the wire. When the material penetrating the belt catches the wire, the wire is pulled off or the tension of the wire is increased, so that the corresponding limit switch acts, and the conveyor is controlled to stop.
2. Leak detector
The detector consists of a tray, a fulcrum, a balance weight, a switch and the like. When the belt is longitudinally torn, materials on the belt leak into the tray through the crack, the weight of the materials overcomes the weight of the counter weight, the whole device rotates around the pivot, the limit switch is forced to act, and then the conveyor is stopped. The detector has simple structure and convenient detection. However, when the conveyor belt is torn, the detector can only detect if there is material on the conveyor belt and the conveyor belt is sufficiently torn to leak material. And when dust on the carrier roller gathers more, misoperation can be caused.
3. Computer vision based deep learning detection
Deep learning is a machine learning technique that can train a model to learn tear features in a belt image. This method requires a large amount of labeled data, i.e., a belt image known to be torn or not. Once the model training is complete, it can be used to detect tears in the new belt image.
4. Visual inspection based on line laser and CCD camera
The visual detection based on line laser and CCD camera is a common three-dimensional visual detection method, which acquires three-dimensional information of the object surface by projecting line laser to the object surface and then capturing the projection of the laser line on the object surface by the CCD camera. For belt longitudinal tear detection, this method can provide more detailed and accurate belt surface information, thereby improving detection accuracy.
In view of the above, the invention provides a belt longitudinal tearing detection method based on two-dimensional point cloud identification.
Disclosure of Invention
(one) solving the technical problems
Aiming at the defects of the prior art, the invention provides a belt longitudinal tearing detection method based on multi-frame two-dimensional point cloud identification.
(II) technical scheme
In order to achieve the above purpose, the present invention provides the following technical solutions: a belt longitudinal tearing detection method based on multi-frame two-dimensional point cloud identification comprises neural network model training and neural network model prediction;
the neural network model training comprises the following steps:
s1: preparing data required for training;
s2: building a neural network to be trained;
s3: preprocessing the data in the step S1;
s4: training a neural network model;
s5: training iteration and loss calculation;
s6: obtaining a network model weight;
the neural network model prediction comprises the following steps:
k1: two-dimensional point cloud;
k2: reading a point cloud frame;
k3: running neural network prediction;
and K4: judging whether the point cloud is abnormal or not;
and K5: judging whether the number of frames exceeds three;
k6: alarming and uploading;
the obtaining of the network model weight in S6 also requires network neural prediction.
Preferably, the method for detecting the longitudinal tearing of the belt based on multi-frame two-dimensional point cloud recognition is provided, the belt tearing detection device based on two-dimensional point cloud recognition is mainly composed of a core switch of a patrol front-end acquisition device and a patrol terminal, the data images of laser vision detection and reflection are compared in real time, the characteristics of the torn part after the belt tearing are analyzed by utilizing multi-point data through a rear-end PointNet neural network intelligent recognition algorithm, and the patrol terminal mainly comprises an algorithm server;
the belt tearing detection device is arranged at the lower part of an upper layer belt of the belt conveyor, is arranged in the running direction of the belt of the feeding port within 5-10 meters, is fixed on a main support of the belt conveyor through a mounting support and is positioned between the upper belt and a lower belt, and can monitor a non-working surface of the belt, identify and judge whether tearing abnormality occurs in real time; the line laser vision and laser camera detects the non-working surface of the belt, the vision captures the camera, and the edge calculation host judges whether the belt is torn or not through the reflected image and the vision feedback, so that a shutdown signal is transmitted to the belt control system to stop the belt in an emergency in the shortest time.
Preferably, in the step S3, the data preprocessing includes:
denoising: removing noise in the point cloud data by using a filtering algorithm;
downsampling: if the point cloud density is high, a downsampling algorithm can be used to reduce the number of point clouds;
normalizing: the coordinate range of the point cloud data is normalized to an appropriate range to facilitate input of the deep learning model.
Preferably, the filtering algorithm includes gaussian filtering and median filtering.
Preferably, the downsampling algorithm comprises voxel grid sampling.
Preferably, the K1 includes: only x-axis and z-axis coordinates are retained where each grid cell represents a voxel and point cloud information within that voxel is recorded.
Preferably, the point cloud data set is divided into a training set, a verification set and a test set, wherein 70% of data is generally used as the training set, 15% of data is used as the verification set, and 15% of data is used as the test set.
Preferably, the step S4 includes: the deep learning model is trained using a training set. The cross entropy loss function may be used to optimize the model parameters.
(III) beneficial effects
Compared with the prior art, the invention provides a belt longitudinal tearing detection method based on multi-frame two-dimensional point cloud identification, which has the following beneficial effects:
the belt longitudinal tearing detection method based on multi-frame two-dimensional point cloud identification is high in accuracy, and the laser radar is high in ranging accuracy, can accurately detect surface variation and irregularity of the belt, and is beneficial to timely finding out the phenomenon of belt tearing. The sensitivity is high, and the laser radar can identify tiny distance changes, so that detection and early warning can be performed even in the initial stage of belt tearing. The detection effect of the laser radar with high reliability is not influenced by ambient light, dust, moisture and the like, so that reliable detection can be realized in various complex environments. The laser radar is non-contact detection equipment, does not cause extra physical damage to the detected object, and is beneficial to prolonging the service life of the belt. Through automatic data acquisition and analysis of the laser radar, the detection efficiency can be greatly improved, and the input of human resources is reduced. The use cost of the user is reduced, and the purposes of reducing the cost, enhancing the efficiency and improving the safety are achieved.
Drawings
The accompanying drawings are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate the invention and together with the embodiments of the invention, serve to explain the invention. In the drawings:
FIG. 1 is a schematic view of the apparatus of the present invention;
FIG. 2 is a flow chart 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.
Examples of the embodiments are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to like or similar elements throughout or elements having like or similar functionality. The embodiments described below by referring to the drawings are illustrative and intended to explain the present invention and should not be construed as limiting the invention.
In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings are merely for convenience in describing the present invention and simplifying the description, and do not indicate or imply that the device or element being referred to must have a specific orientation, be configured and operated in a specific orientation, and therefore should not be construed as limiting the present invention.
In the present invention, unless explicitly specified and limited otherwise, the terms "mounted," "connected," "secured," and the like are to be construed broadly, and may be, for example, fixedly connected, detachably connected, or integrally formed; can be mechanically or electrically connected; can be directly connected or indirectly connected through an intermediate medium, and can be communicated with the inside of two elements or the interaction relationship of the two elements. The specific meaning of the above terms in the present invention can be understood by those of ordinary skill in the art according to the specific circumstances.
1-2, the invention provides a belt longitudinal tearing detection method based on multi-frame two-dimensional point cloud recognition, which comprises neural network model training and neural network model prediction;
the neural network model training comprises the following steps:
s1: preparing data required for training;
s2: building a neural network to be trained;
s3: preprocessing the data in the step S1;
s4: training a neural network model;
s5: training iteration and loss calculation;
s6: obtaining a network model weight;
the neural network model prediction comprises the following steps:
k1: two-dimensional point cloud;
k2: reading a point cloud frame;
k3: running neural network prediction;
and K4: judging whether the point cloud is abnormal or not;
and K5: judging whether the number of frames exceeds three;
k6: alarming and uploading;
the obtaining of the network model weights in S6 also requires network neural prediction.
The belt longitudinal tearing detection device based on two-dimensional point cloud identification is mainly composed of a core switch of a patrol front-end acquisition device and a patrol terminal, wherein laser vision detection and reflected data images are compared in real time, multipoint data are utilized, a rear-end PointNet neural network intelligent recognition algorithm is utilized to analyze the generation characteristics of a torn part after belt tearing, and the patrol terminal mainly comprises an algorithm server;
the belt tearing detection device is arranged at the lower part of an upper layer belt of the belt conveyor, is arranged in the running direction of the belt of the feeding port within 5-10 meters, is fixed on a main support of the belt conveyor through a mounting support and is positioned between the upper belt and a lower belt, and can monitor a non-working surface of the belt, and can identify and judge whether tearing abnormality occurs in real time; the line laser vision and laser camera detects the non-working surface of the belt, the vision captures the camera, and the edge calculation host judges whether the belt is torn or not through the reflected image and the vision feedback, so that a shutdown signal is transmitted to the belt control system to stop the belt in an emergency in the shortest time.
In S3, the data preprocessing includes:
denoising: removing noise in the point cloud data by using a filtering algorithm;
downsampling: if the point cloud density is high, a downsampling algorithm can be used to reduce the number of point clouds;
normalizing: the coordinate range of the point cloud data is normalized to an appropriate range to facilitate input of the deep learning model.
The filtering algorithm includes gaussian filtering and median filtering.
The downsampling algorithm includes voxel grid sampling.
K1 comprises: only x-axis and z-axis coordinates are retained where each grid cell represents a voxel and point cloud information within that voxel is recorded.
The point cloud data set is divided into a training set, a verification set and a test set, wherein 70% of data is generally used as the training set, 15% of data is used as the verification set, and 15% of data is used as the test set.
S4 comprises the following steps: the deep learning model is trained using a training set. The cross entropy loss function may be used to optimize the model parameters.
By constructing a deep learning model: the model PointNet model based on the convolutional neural network is selected, so that point cloud data can be directly processed, and characteristics can be extracted from the point cloud data;
by tear detection: by identifying the data image of the belt, anomalies in the surface of the belt, such as flatness, width, direction of movement, etc., can be found. When the detected data abnormality exceeds a set threshold, the belt is determined to be torn.
It is noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising a reference structure" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
Claims (8)
1. A belt longitudinal tearing detection method based on multi-frame two-dimensional point cloud identification is characterized by comprising the following steps of: the method comprises neural network model training and neural network model prediction;
the neural network model training comprises the following steps:
s1: preparing data required for training;
s2: building a neural network to be trained;
s3: preprocessing the data in the step S1;
s4: training a neural network model;
s5: training iteration and loss calculation;
s6: obtaining a network model weight;
the neural network model prediction comprises the following steps:
k1: two-dimensional point cloud;
k2: reading a point cloud frame;
k3: running neural network prediction;
and K4: judging whether the point cloud is abnormal or not;
and K5: judging whether the number of frames exceeds three;
k6: alarming and uploading;
the obtaining of the network model weight in S6 also requires network neural prediction.
2. The method for detecting longitudinal tearing of the belt based on multi-frame two-dimensional point cloud identification according to claim 1, wherein the method comprises the following steps of: according to the belt longitudinal tearing detection method based on multi-frame two-dimensional point cloud identification, a belt tearing detection device based on two-dimensional point cloud identification is provided, the belt tearing detection device mainly comprises a core switch of a patrol front-end acquisition device and a patrol terminal, a data image of laser visual detection and reflection is compared in real time, multi-point data is utilized, a rear-end PointNet neural network intelligent identification algorithm is utilized, characteristics of a torn part after belt tearing are analyzed, and the patrol terminal mainly comprises an algorithm server;
the belt tearing detection device is arranged at the lower part of an upper layer belt of the belt conveyor, is arranged in the running direction of the belt of the feeding port within 5-10 meters, is fixed on a main support of the belt conveyor through a mounting support and is positioned between the upper belt and a lower belt, and can monitor a non-working surface of the belt, identify and judge whether tearing abnormality occurs in real time; the line laser vision and laser camera detects the non-working surface of the belt, the vision captures the camera, and the edge calculation host judges whether the belt is torn or not through the reflected image and the vision feedback, so that a shutdown signal is transmitted to the belt control system to stop the belt in an emergency in the shortest time.
3. The method for detecting longitudinal tearing of the belt based on multi-frame two-dimensional point cloud identification according to claim 1, wherein the method comprises the following steps of: in the step S3, the data preprocessing includes:
denoising: removing noise in the point cloud data by using a filtering algorithm;
downsampling: if the point cloud density is high, a downsampling algorithm can be used to reduce the number of point clouds;
normalizing: the coordinate range of the point cloud data is normalized to an appropriate range to facilitate input of the deep learning model.
4. The method for detecting longitudinal tearing of the belt based on multi-frame two-dimensional point cloud identification according to claim 1, wherein the method comprises the following steps of: the filtering algorithm includes gaussian filtering and median filtering.
5. The method for detecting longitudinal tearing of the belt based on multi-frame two-dimensional point cloud identification according to claim 1, wherein the method comprises the following steps of: the downsampling algorithm includes a voxel grid sampling method.
6. The method for detecting longitudinal tearing of the belt based on multi-frame two-dimensional point cloud identification according to claim 1, wherein the method comprises the following steps of: the K1 comprises: only x-axis and z-axis coordinates are retained where each grid cell represents a voxel and point cloud information within that voxel is recorded.
7. The method for detecting longitudinal tearing of the belt based on multi-frame two-dimensional point cloud identification according to claim 1, wherein the method comprises the following steps of: the point cloud data set is divided into a training set, a verification set and a test set, wherein 70% of data is generally adopted as the training set, 15% of data is adopted as the verification set, and 15% of data is adopted as the test set.
8. The method for detecting longitudinal tearing of the belt based on multi-frame two-dimensional point cloud identification according to claim 1, wherein the method comprises the following steps of: the step S4 includes: the deep learning model is trained using a training set. The cross entropy loss function may be used to optimize the model parameters.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311180502.2A CN117197743A (en) | 2023-09-13 | 2023-09-13 | Belt longitudinal tearing detection method based on multi-frame two-dimensional point cloud identification |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311180502.2A CN117197743A (en) | 2023-09-13 | 2023-09-13 | Belt longitudinal tearing detection method based on multi-frame two-dimensional point cloud identification |
Publications (1)
Publication Number | Publication Date |
---|---|
CN117197743A true CN117197743A (en) | 2023-12-08 |
Family
ID=88984654
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202311180502.2A Pending CN117197743A (en) | 2023-09-13 | 2023-09-13 | Belt longitudinal tearing detection method based on multi-frame two-dimensional point cloud identification |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117197743A (en) |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20180268256A1 (en) * | 2017-03-16 | 2018-09-20 | Aquifi, Inc. | Systems and methods for keypoint detection with convolutional neural networks |
CN111444811A (en) * | 2020-03-23 | 2020-07-24 | 复旦大学 | Method for detecting three-dimensional point cloud target |
CN111860060A (en) * | 2019-04-29 | 2020-10-30 | 顺丰科技有限公司 | Target detection method and device, terminal equipment and computer readable storage medium |
US20210019891A1 (en) * | 2019-07-15 | 2021-01-21 | Mujin, Inc. | System and method of object detection based on image data |
CN113320924A (en) * | 2021-05-17 | 2021-08-31 | 中国矿业大学(北京) | Belt longitudinal tearing detection device based on single line laser radar |
CN114359246A (en) * | 2022-01-10 | 2022-04-15 | 南京中远通科技有限公司 | Conveyor belt detection method, device, system, electronic device and medium |
CN115027904A (en) * | 2022-06-01 | 2022-09-09 | 南京北新智能科技有限公司 | Belt tearing detection method based on example segmentation |
-
2023
- 2023-09-13 CN CN202311180502.2A patent/CN117197743A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20180268256A1 (en) * | 2017-03-16 | 2018-09-20 | Aquifi, Inc. | Systems and methods for keypoint detection with convolutional neural networks |
CN111860060A (en) * | 2019-04-29 | 2020-10-30 | 顺丰科技有限公司 | Target detection method and device, terminal equipment and computer readable storage medium |
US20210019891A1 (en) * | 2019-07-15 | 2021-01-21 | Mujin, Inc. | System and method of object detection based on image data |
CN111444811A (en) * | 2020-03-23 | 2020-07-24 | 复旦大学 | Method for detecting three-dimensional point cloud target |
CN113320924A (en) * | 2021-05-17 | 2021-08-31 | 中国矿业大学(北京) | Belt longitudinal tearing detection device based on single line laser radar |
CN114359246A (en) * | 2022-01-10 | 2022-04-15 | 南京中远通科技有限公司 | Conveyor belt detection method, device, system, electronic device and medium |
CN115027904A (en) * | 2022-06-01 | 2022-09-09 | 南京北新智能科技有限公司 | Belt tearing detection method based on example segmentation |
Non-Patent Citations (1)
Title |
---|
吴胜昔;卢文建;刘威;顾幸生;: "物体三维断层点云重建的光幕测量方法", 华东理工大学学报(自然科学版), no. 05, 30 October 2018 (2018-10-30) * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10942133B2 (en) | Method and system for automatically detecting fabric defect, and computer readable storage medium | |
CN102221556B (en) | Device and method for online detection on appearance defect of minitype connecting part based on machine vision | |
CN109230351A (en) | A kind of recognition methods that belt conveyor is operating abnormally | |
CN109489724B (en) | Tunnel train safe operation environment comprehensive detection device and detection method | |
CN104608799A (en) | Information fusion technology based train wheel set tread damage online detection and recognition method | |
CN108711148A (en) | A kind of wheel tyre defect intelligent detecting method based on deep learning | |
CN109969736A (en) | A kind of large size carrier strip deviation fault intelligent detecting method | |
CN103837087B (en) | Pantograph automatic testing method based on active shape model | |
CN104535356A (en) | Method and system for monitoring rope arrangement faults of drum steel wire rope on basis of machine vision | |
CN102175692A (en) | System and method for detecting defects of fabric gray cloth quickly | |
CN111127448B (en) | Method for detecting air spring fault based on isolated forest | |
CN110186375A (en) | Intelligent high-speed rail white body assemble welding feature detection device and detection method | |
CN111591715A (en) | Belt longitudinal tearing detection method and device | |
CN110631832A (en) | Subway vehicle bearing fault on-line detection method | |
CN105809219B (en) | A kind of the prefabricated pipe section quality testing statistical system and method for tunnel prefabricated pipe section production line | |
CN111565024A (en) | Solar cell panel fault detection and operation efficiency prediction device and method | |
CN117197743A (en) | Belt longitudinal tearing detection method based on multi-frame two-dimensional point cloud identification | |
CN114324382A (en) | Panel terminal cleanliness detection method and panel terminal cleanliness detection device | |
CN111767913A (en) | Motor train unit vehicle fault dynamic image detection method based on deep learning | |
CN107194923B (en) | Ultraviolet image diagnosis method for defect inspection of contact network power equipment | |
CN116448764A (en) | Automatic crack detection method for fatigue test of aircraft structure | |
CN110376211A (en) | A kind of the synthetic leather crimping on-line measuring device and method of wet process gluing | |
CN115187880A (en) | Communication optical cable defect detection method and system based on image recognition and storage medium | |
Yu et al. | Surface defect detection of hight-speed railway hub based on improved YOLOv3 algorithm | |
US20230169642A1 (en) | Inspecting Sheet Goods Using Deep Learning |
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 |