CN112040174A - Underground coal flow visual detection method - Google Patents
Underground coal flow visual detection method Download PDFInfo
- Publication number
- CN112040174A CN112040174A CN202010697817.4A CN202010697817A CN112040174A CN 112040174 A CN112040174 A CN 112040174A CN 202010697817 A CN202010697817 A CN 202010697817A CN 112040174 A CN112040174 A CN 112040174A
- Authority
- CN
- China
- Prior art keywords
- image
- coal flow
- images
- data set
- coal
- 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
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D21/00—Measuring or testing not otherwise provided for
- G01D21/02—Measuring two or more variables by means not covered by a single other subclass
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a visual detection method for underground coal flow, which comprises the following steps of 1, adjusting the angle of an original camera, and reading a real-time video stream; step 2, based on the video stream read in the step 1, reducing the image detection range in a mode of extracting key frames and dividing image feature sensitive areas to obtain a normalized key frame image; step 3, preprocessing the image obtained in the step 2 by adopting a mode of extracting a reflection component of the key frame image; step 4, acquiring a coal flow data image, preprocessing the data image and establishing a coal flow classification data set to obtain a coal flow image classification model; and 5, reading the image preprocessed in the step 3 and substituting the image into the coal flow image classification model obtained in the step 4 to obtain and output the coal flow classification of the real-time image. The invention solves the problem of low detection precision caused by the existing detection method only depending on the color characteristics of the image.
Description
Technical Field
The invention belongs to the technical field of visual detection, and relates to a visual detection method for underground coal flow.
Background
With the continuous large-scale and intelligent development of underground belt conveying systems, the length of the existing underground belt conveyor can reach hundreds of meters or even thousands of meters, and the load of the conveying system generally accounts for 30% of the total load of coal mine enterprises. In actual production, coal mine enterprises usually keep the high-speed transportation state of the belt at 3-5m/s, and the total load and underground mechanical loss of the coal mine enterprises are undoubtedly increased. Therefore, the intelligent speed regulation of the belt becomes one of main means for reducing energy consumption, and the real-time belt coal flow detection in the intelligent speed regulation becomes the key point for controlling the belt speed. How to green detect the coal flow, avoid the problem of secondary energy waste and the like of coal mine enterprises caused by high-speed belt running under the condition of no coal or less coal, thereby reducing the production load of the coal mine enterprises and promoting the development process of intelligent mines in China.
Along with the development of image processing technology, the coal flow rate of image detection is convenient to install and low in cost, and a new research idea is provided for coal flow rate detection. Chinese patent (CN201810591811.1) discloses a coal flow monitoring method based on image information acquisition, which mainly sets a color threshold value, converts an original image into a binary image, detects the coal flow by using color characteristics, but is easily influenced by the reflection of a material block and a belt in the coal flow; chinese patent (CN201810151060.1) discloses an underground coal flow detection method based on image recognition, which utilizes two-frame difference of a moving coal flow to calculate the coal flow, but when the coal flow is subjected to wet dust removal, the color characteristics of the coal flow of the front frame and the rear frame are too similar, and the moving characteristics are difficult to distinguish. For this reason, it is necessary to propose further improvements.
Disclosure of Invention
The invention aims to provide a visual detection method for underground coal flow, which solves the problem of low detection precision caused by the fact that the existing detection method only depends on image color characteristics or frame difference characteristics.
The invention adopts the technical scheme that the visual detection method for the underground coal flow specifically comprises the following steps:
step 1, adjusting the angle of an original camera, and reading a real-time video stream;
step 2, based on the video stream read in the step 1, reducing the image detection range in a mode of extracting key frames and dividing image feature sensitive areas to obtain a normalized key frame image;
step 3, preprocessing the image obtained in the step 2 by adopting a mode of extracting a reflection component of the key frame image;
step 4, acquiring a coal flow data image, preprocessing the data image, establishing a coal flow classification data set, reading a coal flow classification data training set, bringing the coal flow classification data training set into an improved ResNet network for training, and obtaining a coal flow image classification model;
and 5, reading the image preprocessed in the step 3 and substituting the image into the coal flow image classification model obtained in the step 4 to obtain and output the coal flow classification of the real-time image.
The present invention is also characterized in that,
in step 1, on the basis that the camera is perpendicular to the belt, the angle of the camera is adjusted by 0-15 degrees.
The specific process of the step 2 is as follows:
step 2.1, adjusting the speed of 24 frames/s of the original video stream to the reading speed of 0.5 frames/s to obtain an input video stream;
step 2.2, taking a coal flow belt in an input video stream as an image sensitive area, and dividing the ROI sensitive area to obtain a coal flow belt sensitive area image; and the coal flow belt sensitive area image is normalized to 224 x 224 pixel size, resulting in a normalized key frame image.
In step 3, extracting a reflection component of the key frame image, wherein the extraction of the reflection component is shown in the following formula (1):
wherein I (x, y) is the normalized key frame image, R (x, y) is the reflection component of the normalized key frame image,
the maximum pixel value under the red, green and blue channels; and obtaining a preprocessed image with enhanced image brightness and contrast.
The specific process of the step 4 is as follows:
step 4.1, acquiring historical underground coal flow video images, preprocessing the coal flow images and constructing a coal flow classification data set;
step 4.2, correcting the training image category by using a cross entropy loss function, wherein the cross entropy loss function is shown in the following formula (2):
where C is the number of classes present in the data, p (k) is the probability of the kth class in the predicted image,
probability of k category in actual image;
step 4.3, adjusting the ResNet network model learning rate by using an Adam optimization function, setting the initial learning rate to be 0.001, and adjusting the number of learning rounds to be 200;
step 4.4, obtaining an improved ResNet network based on the image type corrected in the step 4.2 and the ResNet network model learning rate parameter adjusted in the step 4.3;
and 4.5, extracting the coal flow classification data set obtained in the step 4.1, and carrying out training by introducing the coal flow classification data set into an improved ResNet network to obtain a trained coal flow classification model.
The specific process of the step 4.1 is as follows:
step 4.1.1, acquiring historical underground coal flow video images;
step 4.1.2, frame splitting is carried out on the historical video at the rate of 10 s/piece, and historical underground coal flow image data are obtained;
step 4.1.3, dividing the ROI sensitive area of the underground coal flow image by taking the belt coal flow as a sensitive area, and normalizing the divided sensitive area image into a normalized image with the size of 224 × 224 pixels;
step 4.1.4, extracting image reflection components of all the normalized images through a formula (1), and enhancing the brightness and contrast of the images; extracting 50% of images in the reflection component images, turning and performing Gamma enhancement to finish the data preprocessing, and obtaining coal flow data consisting of the reflection component images and the preprocessed images:
wherein I (x, y) is coalA stream image, R (x, y) is a pre-processed image,
the maximum pixel value under the red, green and blue channels; and obtaining a preprocessed image.
And 4.1.5, dividing all the data set images obtained in the step 4.1.4 into three types, namely 10%, 50% and 90%, establishing a coal flow classification data set, and taking 80% of the coal flow classification data set as a model training data set and 20% of the coal flow classification data set as a test data set.
6. In step 4.1.4, 50% of the images are inverted and Gamma-enhanced, wherein 25% of the images are inverted, 25% of the images are subjected to Gamma-enhanced image data with power parameter of 0.5 and parameter of 10.
The invention has the following beneficial effects:
(1) the method is characterized in that a coal flow data set most suitable for local coal mine enterprises is established, and because China is wide in territory, coal rock characteristics of the fully-mechanized mining face of each local coal mine enterprise are different, original video resources of the industrial and mining enterprises are fully and effectively utilized, and resource cost is reduced. Frame dismantling and normalization are carried out according to historical coal flow images, and a three-classification coal flow database is established, so that the method has irreplaceability and non-transportability;
(2) modifying the training image category by using a cross entropy loss function through a training improved ResNet network;
(3) the original image is preprocessed in an image enhancement mode, the reflection component of the key frame image is extracted, the preprocessed image with enhanced image brightness and contrast is obtained, and the influence of underground illumination non-uniformity on extraction of coal flow characteristics is avoided.
Drawings
FIG. 1 is a flow chart of a method of visual inspection of coal flow downhole according to the present invention.
Detailed Description
The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.
The invention discloses a visual detection method for underground coal flow, which has a flow as shown in figure 1 and specifically comprises the following steps:
step 1, adjusting the angle of an original camera, and reading a real-time video stream;
the original camera pixels are 200 thousands, the original size of the real-time video stream is 1920 x 1080 pixels, the camera is vertically spaced from the belt by 1.5m, and the left edge and the right edge of the belt in the image are approximately kept horizontal with the left edge and the right edge of the image when the angle of the camera is adjusted; on camera and the perpendicular basis of belt, with the angular adjustment 0 ~ 15 of camera.
Step 2, based on the video stream read in the step 1, reducing the image detection range in a mode of extracting key frames and dividing image feature sensitive areas to obtain a normalized key frame image;
the specific process of the step 2 is as follows:
step 2.1, adjusting the speed of 24 frames/s of the original video stream to the reading speed of 0.5 frames/s to obtain an input video stream;
step 2.2, taking a coal flow belt in an input video stream as an image sensitive area, and dividing an ROI (region of interest) sensitive area to obtain an image of the coal flow belt sensitive area; and the coal flow belt sensitive area image is normalized to 224 x 224 pixel size, resulting in a normalized key frame image. The ROI sensitive area is divided into quadrangles, and the four-side angular points of the ROI sensitive area are four vertexes of the upper edge of the belt and the lower edge of the belt in the image respectively;
step 3, for the underground images with uneven brightness and low contrast, preprocessing the images obtained in the step 2 in a manner of extracting the reflection components of the key frame images;
in step 3, extracting a reflection component of the key frame image, wherein the extraction of the reflection component is shown in the following formula (1):
wherein I (x, y) is the normalized key frame image, R (x, y) is the reflection component of the normalized key frame image,
under the three channels of red, green and blueA maximum pixel value; and obtaining a preprocessed image with enhanced image brightness and contrast.
Step 4, acquiring a coal flow data image, preprocessing the data image, establishing a coal flow classification data set, reading a coal flow classification data training set, bringing the coal flow classification data training set into an improved ResNet network for training, and obtaining a coal flow image classification model;
the specific process of the step 4 is as follows:
step 4.1, acquiring historical underground coal flow video images, preprocessing the coal flow images and constructing a coal flow classification data set;
the specific process of the step 4.1 is as follows:
step 4.1.1, acquiring historical underground coal flow video images;
step 4.1.2, frame splitting is carried out on the historical video at the rate of 10 s/piece, and historical underground coal flow image data are obtained;
step 4.1.3, dividing the ROI sensitive area of the underground coal flow image by taking the belt coal flow as a sensitive area, and normalizing the divided sensitive area image into a normalized image with the size of 224 × 224 pixels;
step 4.1.4, extracting image reflection components of all the normalized images through a formula (1), and enhancing the brightness and contrast of the images; extracting 50% of images in the reflection component images, turning and performing Gamma enhancement to finish the data preprocessing, and obtaining coal flow data consisting of the reflection component images and the preprocessed images:
wherein I (x, y) is a coal flow image, R (x, y) is a preprocessed image,
the maximum pixel value under the red, green and blue channels; and obtaining a preprocessed image.
Step 4.1.4, overturning and Gamma enhancing 50% of the images, wherein 25% of the images are overturned, the 25% of the images pass through Gamma enhanced image data with power parameter of 0.5 and parameter of 10;
and 4.1.5, dividing all the data set images obtained in the step 4.1.4 into three types, namely 10%, 50% and 90%, establishing a coal flow classification data set, and taking 80% of the coal flow classification data set as a model training data set and 20% of the coal flow classification data set as a test data set. The lump data of the coal flow image classification data is more than 2000 sheets at least.
Step 4.2, correcting the training image category by using a cross entropy loss function, wherein the cross entropy loss function is shown in the following formula (2):
where C is the number of classes present in the data, p (k) is the probability of the kth class in the predicted image,
probability of k category in actual image;
step 4.3, adjusting the ResNet network model learning rate by using an Adam optimization function, setting the initial learning rate to be 0.001, and adjusting the number of learning rounds to be 200;
step 4.4, obtaining an improved ResNet network based on the image type corrected in the step 4.2 and the ResNet network model learning rate parameter adjusted in the step 4.3;
and 4.5, extracting the coal flow classification data set obtained in the step 4.1, and carrying out training by introducing the coal flow classification data set into an improved ResNet network to obtain a trained coal flow classification model.
And 5, reading the image preprocessed in the step 3 and substituting the image into the coal flow image classification model obtained in the step 4 to obtain and output the coal flow classification of the real-time image.
If the real-time image belongs to the 10% category, slowing the running speed of the belt to 1-1.5m/s and keeping the belt running at the speed; if the real-time image belongs to the 50% category, the belt is operated at a constant speed of 2.5-3.5 m/s; if the real-time image belongs to the 90% category, the belt speed is increased to 4-5m/s and the belt is kept running at that speed.
Taking a belt transportation image of a flat coal four-mine underground as an example, an Adam optimization function is selected, and before a loss function is not modified, the image classification detection precision can reach 85.1%; compare original classification precision and can promote 6.3% after being revised as cross entropy with original loss function, classification precision reaches 91.4% to satisfy real-time demand, can pass through image detection mode real-time detection coal flow size, thereby control belt speed in the pit, reduce unnecessary energy loss in the production of colliery enterprise.
Claims (7)
1. A visual detection method for underground coal flow is characterized by comprising the following steps: the method specifically comprises the following steps:
step 1, adjusting the angle of an original camera, and reading a real-time video stream;
step 2, based on the video stream read in the step 1, reducing the image detection range in a mode of extracting key frames and dividing image feature sensitive areas to obtain a normalized key frame image;
step 3, preprocessing the image obtained in the step 2 by adopting a mode of extracting a reflection component of the key frame image;
step 4, acquiring a coal flow data image, preprocessing the data image, establishing a coal flow classification data set, reading a coal flow classification data training set, bringing the coal flow classification data training set into an improved ResNet network for training, and obtaining a coal flow image classification model;
and 5, reading the image preprocessed in the step 3 and substituting the image into the coal flow image classification model obtained in the step 4 to obtain and output the coal flow classification of the real-time image.
2. The visual inspection method of coal flow rate for downhole according to claim 1, wherein: in step 1, on the basis that the camera is perpendicular to the belt, the angle of the camera is adjusted by 0-15 degrees.
3. The visual inspection method of coal flow rate for downhole according to claim 1, wherein: the specific process of the step 2 is as follows:
step 2.1, adjusting the speed of 24 frames/s of the original video stream to the reading speed of 0.5 frames/s to obtain an input video stream;
step 2.2, taking a coal flow belt in an input video stream as an image sensitive area, and dividing the ROI sensitive area to obtain a coal flow belt sensitive area image; and the coal flow belt sensitive area image is normalized to 224 x 224 pixel size, resulting in a normalized key frame image.
4. A method of visual inspection of coal flow downhole according to claim 3, wherein: in step 3, the reflection component of the key frame image is extracted, and the extraction of the reflection component is shown in the following formula (1):
wherein I (x, y) is the normalized key frame image, R (x, y) is the reflection component of the normalized key frame image,
the maximum pixel value under the red, green and blue channels; and obtaining a preprocessed image with enhanced image brightness and contrast.
5. The visual inspection method of coal flow rate for downhole according to claim 4, wherein: the specific process of the step 4 is as follows:
step 4.1, acquiring historical underground coal flow video images, preprocessing the coal flow images and constructing a coal flow classification data set;
step 4.2, correcting the training image category by using a cross entropy loss function, wherein the cross entropy loss function is shown in the following formula (2):
wherein C is a class present in the dataNumber of classes, p (k) is the probability of the kth class in the predicted image,
probability of k category in actual image;
step 4.3, adjusting the ResNet network model learning rate by using an Adam optimization function, setting the initial learning rate to be 0.001, and adjusting the number of learning rounds to be 200;
step 4.4, obtaining an improved ResNet network based on the image type corrected in the step 4.2 and the ResNet network model learning rate parameter adjusted in the step 4.3;
and 4.5, extracting the coal flow classification data set obtained in the step 4.1, and carrying out training by introducing the coal flow classification data set into an improved ResNet network to obtain a trained coal flow classification model.
6. A visual inspection method of coal flow rate for downhole use according to claim 5, wherein: the specific process of the step 4.1 is as follows:
step 4.1.1, acquiring historical underground coal flow video images;
step 4.1.2, frame splitting is carried out on the historical video at the rate of 10 s/piece, and historical underground coal flow image data are obtained;
step 4.1.3, dividing the ROI sensitive area of the underground coal flow image by taking the belt coal flow as a sensitive area, and normalizing the divided sensitive area image into a normalized image with the size of 224 × 224 pixels;
step 4.1.4, extracting image reflection components of all the normalized images through a formula (1), and enhancing the brightness and contrast of the images; extracting 50% of images in the reflection component images, turning and performing Gamma enhancement to finish the data preprocessing, and obtaining coal flow data consisting of the reflection component images and the preprocessed images:
wherein I (x, y) is a coal flow image, R (x, y) is a preprocessed image,
the maximum pixel value under the red, green and blue channels; and obtaining a preprocessed image.
And 4.1.5, dividing all the data set images obtained in the step 4.1.4 into three types, namely 10%, 50% and 90%, establishing a coal flow classification data set, and taking 80% of the coal flow classification data set as a model training data set and 20% of the coal flow classification data set as a test data set.
7. The visual inspection method of coal flow rate for downhole according to claim 6, wherein: in the step 4.1.4, 50% of the images are inverted and Gamma-enhanced, wherein 25% of the images are inverted, 25% of the images are subjected to Gamma-enhanced image data with power parameter of 0.5 and parameter of 10.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010697817.4A CN112040174A (en) | 2020-07-20 | 2020-07-20 | Underground coal flow visual detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010697817.4A CN112040174A (en) | 2020-07-20 | 2020-07-20 | Underground coal flow visual detection method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112040174A true CN112040174A (en) | 2020-12-04 |
Family
ID=73579333
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010697817.4A Pending CN112040174A (en) | 2020-07-20 | 2020-07-20 | Underground coal flow visual detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112040174A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112561887A (en) * | 2020-12-18 | 2021-03-26 | 中国矿业大学 | Belt conveyor coal flow binocular vision measurement method based on deep migration learning |
| CN113283395A (en) * | 2021-06-28 | 2021-08-20 | 西安科技大学 | Video detection method for blocking foreign matters at coal conveying belt transfer joint |
| CN113536942A (en) * | 2021-06-21 | 2021-10-22 | 上海赫千电子科技有限公司 | Road traffic sign recognition method based on neural network |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107341521A (en) * | 2017-07-10 | 2017-11-10 | 东北大学 | A kind of method based on coal spectroscopic data to grade of coal |
| CN107730473A (en) * | 2017-11-03 | 2018-02-23 | 中国矿业大学 | A kind of underground coal mine image processing method based on deep neural network |
| CN107944394A (en) * | 2017-11-27 | 2018-04-20 | 宁夏广天夏电子科技有限公司 | A kind of video analysis method and system being detected to material in conveyor |
| CN108537192A (en) * | 2018-04-17 | 2018-09-14 | 福州大学 | A kind of remote sensing image ground mulching sorting technique based on full convolutional network |
| CN108664874A (en) * | 2018-02-14 | 2018-10-16 | 北京广天夏科技有限公司 | Underground coal flow rate testing methods based on image recognition |
| CN110675443A (en) * | 2019-09-24 | 2020-01-10 | 西安科技大学 | Coal briquette area detection method for underground coal conveying image |
| CN110675374A (en) * | 2019-09-17 | 2020-01-10 | 电子科技大学 | Two-dimensional image sewage flow detection method based on generation countermeasure network |
| CN110781944A (en) * | 2019-10-21 | 2020-02-11 | 中冶南方（武汉）自动化有限公司 | Automatic molten iron slag-off control method based on deep learning |
-
2020
- 2020-07-20 CN CN202010697817.4A patent/CN112040174A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107341521A (en) * | 2017-07-10 | 2017-11-10 | 东北大学 | A kind of method based on coal spectroscopic data to grade of coal |
| CN107730473A (en) * | 2017-11-03 | 2018-02-23 | 中国矿业大学 | A kind of underground coal mine image processing method based on deep neural network |
| CN107944394A (en) * | 2017-11-27 | 2018-04-20 | 宁夏广天夏电子科技有限公司 | A kind of video analysis method and system being detected to material in conveyor |
| CN108664874A (en) * | 2018-02-14 | 2018-10-16 | 北京广天夏科技有限公司 | Underground coal flow rate testing methods based on image recognition |
| CN108537192A (en) * | 2018-04-17 | 2018-09-14 | 福州大学 | A kind of remote sensing image ground mulching sorting technique based on full convolutional network |
| CN110675374A (en) * | 2019-09-17 | 2020-01-10 | 电子科技大学 | Two-dimensional image sewage flow detection method based on generation countermeasure network |
| CN110675443A (en) * | 2019-09-24 | 2020-01-10 | 西安科技大学 | Coal briquette area detection method for underground coal conveying image |
| CN110781944A (en) * | 2019-10-21 | 2020-02-11 | 中冶南方（武汉）自动化有限公司 | Automatic molten iron slag-off control method based on deep learning |
Non-Patent Citations (1)
| Title |
|---|
| 李宝奇等: "光照不变量特征提取新方法及其在目标识别中的应用", 《电子学报》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112561887A (en) * | 2020-12-18 | 2021-03-26 | 中国矿业大学 | Belt conveyor coal flow binocular vision measurement method based on deep migration learning |
| CN112561887B (en) * | 2020-12-18 | 2021-06-29 | 中国矿业大学 | Belt conveyor coal flow binocular vision measurement method based on deep migration learning |
| CN113536942A (en) * | 2021-06-21 | 2021-10-22 | 上海赫千电子科技有限公司 | Road traffic sign recognition method based on neural network |
| CN113536942B (en) * | 2021-06-21 | 2024-04-12 | 上海赫千电子科技有限公司 | Road traffic sign recognition method based on neural network |
| CN113283395A (en) * | 2021-06-28 | 2021-08-20 | 西安科技大学 | Video detection method for blocking foreign matters at coal conveying belt transfer joint |
| CN113283395B (en) * | 2021-06-28 | 2024-03-29 | 西安科技大学 | Video detection method for blocking foreign matters at transfer position of coal conveying belt |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112040174A (en) | Underground coal flow visual detection method | |
| CN110390691B (en) | Ore dimension measuring method based on deep learning and application system | |
| CN102043950B (en) | Vehicle outline recognition method based on canny operator and marginal point statistic | |
| CN101916364B (en) | Adaptive dial digital identification method | |
| CN111709935B (en) | Real-time coal gangue positioning and identifying method for ground moving belt | |
| CN111161292B (en) | Ore scale measurement method and application system | |
| CN102332094B (en) | Semi-supervised online study face detection method | |
| CN102902974B (en) | Image based method for identifying railway overhead-contact system bolt support identifying information | |
| CN105374015A (en) | Binary method for low-quality document image based on local contract and estimation of stroke width | |
| CN101739549B (en) | Face detection method and system | |
| CN105944976A (en) | Method and device for sorting massive gangue by using digital image processing technology | |
| CN111369538B (en) | Conveyor belt deviation detection method based on deep convolutional neural network | |
| CN106529461A (en) | Vehicle model identifying algorithm based on integral characteristic channel and SVM training device | |
| CN104616006A (en) | Surveillance video oriented bearded face detection method | |
| CN111896541B (en) | Flue gas on-line monitoring system based on block chain | |
| CN112200813A (en) | Coal and gangue identification method and system considering illumination factor | |
| CN112124900B (en) | Visual detection method for underground belt deviation | |
| CN114049589A (en) | Transformer substation monitoring system based on artificial intelligence | |
| CN109271904A (en) | A kind of black smoke vehicle detection method based on pixel adaptivenon-uniform sampling and Bayesian model | |
| CN110519566B (en) | Belt motion state monitoring method based on video processing | |
| CN117437600A (en) | Coal flow monitoring system based on image recognition technology | |
| CN113772364B (en) | Automatic adjusting method of belt conveyor deviation detecting device | |
| CN113658135B (en) | Fuzzy PID-based self-adaptive dimming belt foreign matter detection method and system | |
| CN111573200B (en) | Visual detection system and detection method for blocking of underground belt transfer joint | |
| CN110135274A (en) | A kind of people flow rate statistical method based on recognition of face |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20201204 |