CN112163631A - Gold ore mineral analysis method based on video analysis for orepass - Google Patents
Gold ore mineral analysis method based on video analysis for orepass Download PDFInfo
- Publication number
- CN112163631A CN112163631A CN202011093476.6A CN202011093476A CN112163631A CN 112163631 A CN112163631 A CN 112163631A CN 202011093476 A CN202011093476 A CN 202011093476A CN 112163631 A CN112163631 A CN 112163631A
- Authority
- CN
- China
- Prior art keywords
- ore
- gray level
- image
- occurrence matrix
- neural network
- 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
- 238000004458 analytical method Methods 0.000 title claims abstract description 27
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 title claims abstract description 23
- 239000010931 gold Substances 0.000 title claims abstract description 23
- 229910052737 gold Inorganic materials 0.000 title claims abstract description 23
- 229910052500 inorganic mineral Inorganic materials 0.000 title claims abstract description 21
- 239000011707 mineral Substances 0.000 title claims abstract description 21
- 239000011159 matrix material Substances 0.000 claims abstract description 34
- 238000013528 artificial neural network Methods 0.000 claims abstract description 29
- 238000000034 method Methods 0.000 claims description 20
- 238000004364 calculation method Methods 0.000 claims description 9
- 238000001914 filtration Methods 0.000 claims description 4
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
Abstract
The invention discloses a gold ore mineral analysis method based on video analysis for a drop shaft, which comprises the following steps: step 1, establishing a sample library; step 2, extracting a gray level co-occurrence matrix from the sample image, wherein the gray level co-occurrence matrix is used for recording image texture characteristics; then establishing and training a neural network according to the corresponding relation between the gray level co-occurrence matrix and the ore classification; step 3, acquiring an actual image of the ore to be detected by using a camera, and extracting a gray level co-occurrence matrix from the actual image; and 4, inputting the gray level co-occurrence matrix obtained in the step 3 into a neural network, obtaining the classification of the ore corresponding to each actual image according to the output value of the neural network, and judging the content of the gold ore according to the classification of the ore. The invention can quickly solve the problems of difficult manual mineral distinguishing, strict requirements on the experience of an ore inspector, low efficiency and low accuracy.
Description
Technical Field
The invention relates to a method for analyzing the mineral content of gold ore.
Background
At present, the analysis of the mineral of gold ore mainly depends on special instruments and manual experience, the former has the problems of poor timeliness and high cost, and the latter has the problems of large workload, low efficiency and low accuracy.
Disclosure of Invention
The invention provides a gold ore mineral analysis method based on video analysis for a drop shaft, which aims to: the method has the advantages of quickly analyzing the mineral substances of the gold ores according to the video image data, improving the efficiency, reducing the cost and improving the accuracy.
The technical scheme of the invention is as follows:
a gold ore mineral analysis method based on video analysis at a draw shaft comprises the following steps:
step 1, establishing a sample library, wherein the sample library comprises sample images of ores with different gold ore contents;
step 2, extracting a gray level co-occurrence matrix from the sample image, wherein the gray level co-occurrence matrix is used for recording image texture characteristics; then establishing and training a neural network according to the corresponding relation between the gray level co-occurrence matrix and the ore classification;
step 3, acquiring an actual image of the ore to be detected by using a camera, and extracting a gray level co-occurrence matrix from the actual image;
and 4, inputting the gray level co-occurrence matrix obtained in the step 3 into a neural network, obtaining the classification of the ore corresponding to each actual image according to the output value of the neural network, and judging the content of the gold ore according to the classification of the ore.
As a further improvement of the method: in step 2 and step 3, the step of extracting the gray level co-occurrence matrix is as follows: firstly, shooting a total image, then using a sliding window to intercept a group of local images from the total image as a sample image in the step 2 or an actual image in the step 3, and establishing a gray level co-occurrence matrix according to the arrangement position of each pixel in the local image and the corresponding gray value.
As a further improvement of the method: the neural network is of a structure N1N 2N 3, wherein: n1 is the node number of the input layer, which is equal to the element number of the gray level co-occurrence matrix after being sequentially expanded into a sequence according to rows; n2 is the number of nodes in the middle tier; n3 is the number of nodes of the output level, which is equal to the number of ore classifications;
and sequentially expanding the gray level co-occurrence matrix into a sequence according to rows, then inputting each element of the sequence as an input value of an input layer node into a neural network for calculation, wherein the output value of each output layer node is a reliability value of corresponding classification.
As a further improvement of the method: in step 4, setting to obtain M gray level co-occurrence matrixes, inputting the M gray level co-occurrence matrixes into a neural network to obtain M groups of calculation results, wherein each calculation result comprises N3 credibility values; and respectively adding the M credibility values of the same ore classification to obtain N3 credibility value sums, wherein the ore classification corresponding to the maximum value of the sums is the classification of the ore corresponding to the current actual image.
As a further improvement of the method: in step 2, the gray level co-occurrence matrixes of the sample images are respectively input into the neural network during training, if the ore classification corresponding to the maximum value is consistent with the actual ore classification in the N3 credibility values output by the neural network, the judgment is 'true', otherwise, the judgment is 'false', and the parameter values of the neural network are adjusted according to the judgment result until the training is finished.
As a further improvement of the method: before extracting the gray level co-occurrence matrix, graying and Gaussian filtering are carried out on the image so as to increase the texture information of the image.
As a further improvement of the method: before the image is shot, the surface of the ore is humidified.
As a further improvement of the method: when shooting a sample image and an actual image, light sources are symmetrically arranged on the left side and the right side of the ore.
Compared with the prior art, the invention has the following beneficial effects: (1) based on the characteristic that the ore minerals and the color and the shape of the ore minerals have certain correlation, the method analyzes and processes the ore images through a computer artificial intelligent vision technology, distinguishes different types of ores, and thereby identifies the ore minerals, and can quickly solve the problems that the minerals are difficult to distinguish artificially, strict requirements are required on the experience of an ore inspector, the efficiency is low, and the accuracy is low; (2) according to the invention, the problem of inaccurate judgment result caused by inconsistent arrangement of input sequences of the neural network can be solved by combining the sliding window method and the weighted summation method.
Drawings
FIG. 1 is a schematic illustration of the practice of the present invention.
Detailed Description
The technical scheme of the invention is explained in detail in the following with the accompanying drawings:
as shown in figure 1, the LED light sources are symmetrically arranged on the left side and the right side of the mine car, so that the uniform illumination of the detected ore is ensured. Meanwhile, a high-definition network industrial camera and an industrial personal computer are arranged. The resolution of the camera is 1920 × 1080 lens optical axis and ore horizontal plane are 90。And is located directly above the ore, with its field of view covering the detected ore region. And image recognition software is deployed on the industrial personal computer and is connected with the terminal display. The industrial personal computer and the high-definition network industrial camera are connected in the same local area network.
The implementation steps are as follows:
step 1, establishing a sample library, wherein the sample library comprises sample images of ores with different gold ore contents. Before the image is shot, the surface of the ore is humidified.
Step 2, extracting a gray level co-occurrence matrix from the sample image, wherein the gray level co-occurrence matrix is used for recording image texture characteristics; and then establishing and training a neural network according to the corresponding relation between the gray level co-occurrence matrix and the ore classification.
Specifically, a total image is shot first, and graying and gaussian filtering are performed on the image to increase texture information of the image. And then, a group of local images are intercepted from the total image by using a sliding window to serve as sample images, and a gray level co-occurrence matrix is established according to the arrangement position of each pixel in the local images and the corresponding gray level value. For example, if the resolution of the total image (the region not related to the ore has been removed) is 30 × 40, and the resolution of the sliding window is 10 × 10, the sliding window moves by one pixel at a time, and the number of the total partial images is (30-10) × (40-10) = 600, which corresponds to 600 times of image feature extraction. And then, establishing a gray level co-occurrence matrix according to the arrangement position of each pixel in the local image and the corresponding gray level value.
The neural network is of a structure N1N 2N 3, wherein: n1 is the node number of the input layer, which is equal to the element number of the gray level co-occurrence matrix after being sequentially expanded into a sequence according to rows; n2 is the number of nodes in the middle tier; n3 is the number of nodes in the output level, which is equal to the number of ore classifications, in this example, 3 types of ores in common.
And during training, sequentially expanding the gray level co-occurrence matrix of each sample image into a sequence according to rows, inputting each element of the sequence as an input value of an input layer node into a neural network for calculation, wherein the output value of each output layer node is a reliability value of corresponding classification. And if the ore classification corresponding to the maximum value in the N3 credibility values output by the neural network is consistent with the actual ore classification, judging the ore classification to be true, otherwise, judging the ore classification to be false, adjusting the parameter values of the neural network according to the judgment result until all output results are consistent with the actual results, and finishing the training.
And 3, acquiring an actual image of the ore to be detected by using a camera (as above, intercepting through a sliding window). Before the image is shot, the surface of the ore still needs to be humidified. And then carrying out graying and Gaussian filtering, and extracting a gray level co-occurrence matrix from the actual image.
And 4, inputting the gray level co-occurrence matrix obtained in the step 3 into a trained neural network. Setting to obtain M gray level co-occurrence matrixes, inputting the M gray level co-occurrence matrixes into a neural network to obtain M groups of calculation results, wherein each calculation result comprises N3 credibility values; and respectively adding the M credibility values of the same ore classification to obtain N3 credibility value sums, wherein the ore classification corresponding to the maximum value of the sums is the classification of the ore corresponding to the current actual image. And finally, judging the content of the gold ore according to the ore classification, and outputting and displaying on a terminal display.
Claims (8)
1. A gold ore mineral analysis method based on video analysis at a draw shaft is characterized by comprising the following steps:
step 1, establishing a sample library, wherein the sample library comprises sample images of ores with different gold ore contents;
step 2, extracting a gray level co-occurrence matrix from the sample image, wherein the gray level co-occurrence matrix is used for recording image texture characteristics; then establishing and training a neural network according to the corresponding relation between the gray level co-occurrence matrix and the ore classification;
step 3, acquiring an actual image of the ore to be detected by using a camera, and extracting a gray level co-occurrence matrix from the actual image;
and 4, inputting the gray level co-occurrence matrix obtained in the step 3 into a neural network, obtaining the classification of the ore corresponding to each actual image according to the output value of the neural network, and judging the content of the gold ore according to the classification of the ore.
2. The method for video analysis-based gold mineral analysis at a draw shaft according to claim 1, wherein: in step 2 and step 3, the step of extracting the gray level co-occurrence matrix is as follows: firstly, shooting a total image, then using a sliding window to intercept a group of local images from the total image as a sample image in the step 2 or an actual image in the step 3, and establishing a gray level co-occurrence matrix according to the arrangement position of each pixel in the local image and the corresponding gray value.
3. The method for video analysis-based gold mineral analysis at a draw shaft according to claim 2, wherein: the neural network is of a structure N1N 2N 3, wherein: n1 is the node number of the input layer, which is equal to the element number of the gray level co-occurrence matrix after being sequentially expanded into a sequence according to rows; n2 is the number of nodes in the middle tier; n3 is the number of nodes of the output level, which is equal to the number of ore classifications;
and sequentially expanding the gray level co-occurrence matrix into a sequence according to rows, then inputting each element of the sequence as an input value of an input layer node into a neural network for calculation, wherein the output value of each output layer node is a reliability value of corresponding classification.
4. The method for video analysis-based gold mineral analysis at a draw shaft according to claim 3, wherein: in step 4, setting to obtain M gray level co-occurrence matrixes, inputting the M gray level co-occurrence matrixes into a neural network to obtain M groups of calculation results, wherein each calculation result comprises N3 credibility values; and respectively adding the M credibility values of the same ore classification to obtain N3 credibility value sums, wherein the ore classification corresponding to the maximum value of the sums is the classification of the ore corresponding to the current actual image.
5. The method for video analysis-based gold mineral analysis at a draw shaft according to claim 3, wherein: in step 2, the gray level co-occurrence matrixes of the sample images are respectively input into the neural network during training, if the ore classification corresponding to the maximum value is consistent with the actual ore classification in the N3 credibility values output by the neural network, the judgment is 'true', otherwise, the judgment is 'false', and the parameter values of the neural network are adjusted according to the judgment result until the training is finished.
6. The method for video analysis-based gold mineral analysis at a draw shaft according to claim 1, wherein: before extracting the gray level co-occurrence matrix, graying and Gaussian filtering are carried out on the image so as to increase the texture information of the image.
7. The method for video analysis-based gold mineral analysis at a draw shaft according to claim 1, wherein: before the image is shot, the surface of the ore is humidified.
8. The method for video-based analysis of gold mineral analysis at a draw according to any one of claims 1 to 7, wherein: when shooting a sample image and an actual image, light sources are symmetrically arranged on the left side and the right side of the ore.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011093476.6A CN112163631A (en) | 2020-10-14 | 2020-10-14 | Gold ore mineral analysis method based on video analysis for orepass |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011093476.6A CN112163631A (en) | 2020-10-14 | 2020-10-14 | Gold ore mineral analysis method based on video analysis for orepass |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112163631A true CN112163631A (en) | 2021-01-01 |
Family
ID=73866819
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011093476.6A Pending CN112163631A (en) | 2020-10-14 | 2020-10-14 | Gold ore mineral analysis method based on video analysis for orepass |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112163631A (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106529568A (en) * | 2016-10-11 | 2017-03-22 | 浙江工业大学 | Pearl multi-classification method based on BP neural network |
| CN106920229A (en) * | 2017-01-22 | 2017-07-04 | 北京奇艺世纪科技有限公司 | Image obscuring area automatic testing method and system |
| CN107463167A (en) * | 2016-06-03 | 2017-12-12 | 苏州宝时得电动工具有限公司 | Automatic running device and target area recognition methods |
| US20180012107A1 (en) * | 2015-12-11 | 2018-01-11 | Tencent Technology (Shenzhen) Company Limited | Image classification method, electronic device, and storage medium |
| CN108253938A (en) * | 2017-12-29 | 2018-07-06 | 武汉大学 | TBM broken rock slags digital close range photogrammetry identifies and inversion method |
| CN108805165A (en) * | 2018-04-27 | 2018-11-13 | 淘然视界(杭州)科技有限公司 | A kind of coal identification method for sorting, electronic equipment, storage medium and system |
| US10410092B1 (en) * | 2015-12-16 | 2019-09-10 | Hrl Laboratories, Llc | Automated classification of rock types and analyst-centric visualizations—front end |
| CN110705511A (en) * | 2019-10-16 | 2020-01-17 | 北京字节跳动网络技术有限公司 | Blurred image recognition method, device, equipment and storage medium |
| CN111539356A (en) * | 2020-04-27 | 2020-08-14 | 刘秀萍 | Pavement gap accurate and efficient detection method based on linear array image |
-
2020
- 2020-10-14 CN CN202011093476.6A patent/CN112163631A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180012107A1 (en) * | 2015-12-11 | 2018-01-11 | Tencent Technology (Shenzhen) Company Limited | Image classification method, electronic device, and storage medium |
| US10410092B1 (en) * | 2015-12-16 | 2019-09-10 | Hrl Laboratories, Llc | Automated classification of rock types and analyst-centric visualizations—front end |
| CN107463167A (en) * | 2016-06-03 | 2017-12-12 | 苏州宝时得电动工具有限公司 | Automatic running device and target area recognition methods |
| CN106529568A (en) * | 2016-10-11 | 2017-03-22 | 浙江工业大学 | Pearl multi-classification method based on BP neural network |
| CN106920229A (en) * | 2017-01-22 | 2017-07-04 | 北京奇艺世纪科技有限公司 | Image obscuring area automatic testing method and system |
| CN108253938A (en) * | 2017-12-29 | 2018-07-06 | 武汉大学 | TBM broken rock slags digital close range photogrammetry identifies and inversion method |
| CN108805165A (en) * | 2018-04-27 | 2018-11-13 | 淘然视界(杭州)科技有限公司 | A kind of coal identification method for sorting, electronic equipment, storage medium and system |
| CN110705511A (en) * | 2019-10-16 | 2020-01-17 | 北京字节跳动网络技术有限公司 | Blurred image recognition method, device, equipment and storage medium |
| CN111539356A (en) * | 2020-04-27 | 2020-08-14 | 刘秀萍 | Pavement gap accurate and efficient detection method based on linear array image |
Non-Patent Citations (2)
| Title |
|---|
| 邵珠宏 等: "基于局部特征和集成学习的鲁棒彩色人脸识别算法", 《东南大学学报》 * |
| 高砚军 等: "基于窗口自适应灰度共生矩阵的SAR图像分", 《仪器仪表学报》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111179251B (en) | Defect detection system and method based on twin neural network and by utilizing template comparison | |
| CN110543892B (en) | Part identification method based on multilayer random forest | |
| CN106875373B (en) | Mobile phone screen MURA defect detection method based on convolutional neural network pruning algorithm | |
| CN108717524B (en) | Gesture recognition system based on double-camera mobile phone and artificial intelligence system | |
| CN110852316B (en) | Image tampering detection and positioning method adopting convolution network with dense structure | |
| CN111325203A (en) | American license plate recognition method and system based on image correction | |
| CN113011357B (en) | Depth fake face video positioning method based on space-time fusion | |
| CN108596102B (en) | RGB-D-based indoor scene object segmentation classifier construction method | |
| CN110082821B (en) | Label-frame-free microseism signal detection method and device | |
| CN110738160A (en) | human face quality evaluation method combining with human face detection | |
| CN109948566A (en) | A kind of anti-fraud detection method of double-current face based on weight fusion and feature selecting | |
| CN111428556A (en) | Traffic sign recognition method based on capsule neural network | |
| CN110427888A (en) | A kind of face method for evaluating quality based on feature clustering | |
| CN112487981A (en) | MA-YOLO dynamic gesture rapid recognition method based on two-way segmentation | |
| CN110569780A (en) | high-precision face recognition method based on deep transfer learning | |
| CN110688966B (en) | Semantic guidance pedestrian re-recognition method | |
| CN111160194B (en) | Static gesture image recognition method based on multi-feature fusion | |
| CN110751225A (en) | Image classification method, device and storage medium | |
| CN112529901B (en) | Crack identification method in complex environment | |
| CN112785568A (en) | Magnetic shoe defect segmentation method based on deep learning | |
| CN112861987A (en) | Target detection method under dark light environment | |
| CN113221763A (en) | Flame identification method based on video image brightness | |
| CN116977931A (en) | High-altitude parabolic identification method based on deep learning | |
| CN112163631A (en) | Gold ore mineral analysis method based on video analysis for orepass | |
| CN110633666A (en) | Gesture track recognition method based on finger color patches |
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: 20210101 |