CN113240663A - Conveyor belt ore granularity detection method based on edge response fusion algorithm - Google Patents
Conveyor belt ore granularity detection method based on edge response fusion algorithm Download PDFInfo
- Publication number
- CN113240663A CN113240663A CN202110603450.XA CN202110603450A CN113240663A CN 113240663 A CN113240663 A CN 113240663A CN 202110603450 A CN202110603450 A CN 202110603450A CN 113240663 A CN113240663 A CN 113240663A
- Authority
- CN
- China
- Prior art keywords
- image
- ore
- algorithm
- conveyor belt
- detection method
- 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
- 230000004927 fusion Effects 0.000 title claims abstract description 27
- 230000004044 response Effects 0.000 title claims abstract description 23
- 238000001514 detection method Methods 0.000 title claims abstract description 22
- 238000000034 method Methods 0.000 claims abstract description 37
- 230000009466 transformation Effects 0.000 claims abstract description 16
- 238000003708 edge detection Methods 0.000 claims abstract description 8
- 238000007781 pre-processing Methods 0.000 claims abstract description 5
- 230000000877 morphologic effect Effects 0.000 claims abstract 3
- 230000011218 segmentation Effects 0.000 claims description 28
- 238000001914 filtration Methods 0.000 claims description 9
- 239000002245 particle Substances 0.000 claims description 8
- 230000002146 bilateral effect Effects 0.000 claims description 6
- 238000011176 pooling Methods 0.000 claims description 6
- 239000011800 void material Substances 0.000 claims description 5
- 230000008569 process Effects 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- 238000005457 optimization Methods 0.000 claims description 3
- 230000003044 adaptive effect Effects 0.000 claims description 2
- 230000007797 corrosion Effects 0.000 claims description 2
- 238000005260 corrosion Methods 0.000 claims description 2
- 238000005070 sampling Methods 0.000 claims description 2
- 230000007257 malfunction Effects 0.000 claims 1
- 230000006870 function Effects 0.000 abstract description 8
- 238000004519 manufacturing process Methods 0.000 abstract description 6
- 238000004364 calculation method Methods 0.000 abstract 2
- 238000004458 analytical method Methods 0.000 abstract 1
- 230000001131 transforming effect Effects 0.000 abstract 1
- 230000000694 effects Effects 0.000 description 11
- 238000005065 mining Methods 0.000 description 7
- 238000003709 image segmentation Methods 0.000 description 4
- 238000012360 testing method Methods 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000005422 blasting Methods 0.000 description 2
- 230000000295 complement effect Effects 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 239000003086 colorant Substances 0.000 description 1
- 238000004883 computer application Methods 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000007599 discharging Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 230000003628 erosive effect Effects 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 239000004575 stone Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/12—Edge-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20021—Dividing image into blocks, subimages or windows
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30132—Masonry; Concrete
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Medical Informatics (AREA)
- Quality & Reliability (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention relates to a conveyor belt ore granularity detection method based on an edge response fusion algorithm, which comprises the following steps: (1) shooting an ore image on a conveyor belt in real time by using an industrial camera; (2) preprocessing acquired images by image graying, denoising and image binarization; (3) optimizing the binary image and transforming the distance by using a morphological method; (4) an edge response fusion algorithm combining improved HED network edge detection and a watershed algorithm based on distance transformation is adopted to effectively segment the ore image; (5) calling a part of functions in an OpenCV library to perform granularity analysis calculation on the segmented image; (6) and transmitting the ore granularity information to a crusher control system in real time. The accurate ore granularity obtained by calculation can provide important reference for adjusting related parameters of the crusher, reduce the fault occurrence probability of the ore crusher and improve the production efficiency of mines.
Description
Technical Field
The invention belongs to the technical field of mining, simultaneously belongs to the fields of computer application and deep learning, relates to automatic detection of ore granularity, and particularly relates to a conveyor belt ore granularity detection method based on an edge response fusion algorithm.
Background
In the mining industry, the crushing effect of ores in a crushing station can directly reflect the blasting efficiency. The ore after the explosive blasting is loaded to a truck and then is transported to a crushing station to complete 2 times of crushing, the work is mainly completed by a crusher, and the crushed ore is transported to other stations by a conveyor belt after being crushed by the crusher. The adjustment of the parameters of the crusher is closely related to the size of the ore granularity, and how to determine the size of the ore granularity on the crushed conveying belt so as to properly adjust the parameters of the crusher has important significance for mine production. As a traditional method, manual screening is used in mining areas for many years, so that the efficiency is low, the mining cost is invisibly increased, and the method is not suitable for intelligent development of mines. At present, a plurality of ore granularity detection methods based on vision and image processing technologies are applied to part of mine development so as to adjust the parameters of the crusher to be optimal in time, reduce energy consumption and save cost; bringing considerable economic benefit for mine production.
The use of computer vision and image processing techniques to determine the particle size of ore on a conveyor belt also presents a number of problems in the application process. Due to the complex mine environment, the illumination can not meet the ideal requirement, so that the quality of the obtained ore picture is low and the obtained ore picture is full of noise; and the ore granularity on the conveyer belt is different, and different ores are adhered together, and these all bring huge difficulty to image segmentation. The conventional algorithm based on threshold segmentation has difficulty in obtaining the optimal threshold, which often results in the failure to segment the image to the ideal effect. The watershed algorithm is a classic segmentation algorithm, the problem of adhesion of ores cannot be effectively solved, and over-segmentation and under-segmentation phenomena are easily caused.
Disclosure of Invention
In order to overcome the defects of the prior art and solve the problem of inaccuracy of the existing ore image segmentation algorithm, the invention aims to provide the conveyor belt ore granularity detection method based on the edge response fusion algorithm, the deep learning and the traditional method are combined, the two segmentation results are subjected to one AND operation to exert respective advantages, the respective vacancy of edge detection is filled, the segmentation effect is optimized, and the granularity information of ore is obtained through accurate image segmentation, so that the parameters of a crusher are adjusted, the fault rate of the crusher is reduced, the production efficiency of a mine is improved, and the mining cost is saved.
In order to achieve the purpose, the invention adopts the technical scheme that:
a conveyor belt ore granularity detection method based on an edge response fusion algorithm comprises the following steps:
step (1), using an industrial camera to shoot an ore image on a conveyor belt in real time;
step (2), preprocessing the collected image, including image graying, denoising and image binarization;
step (3), carrying out binarization image optimization and distance transformation on the preprocessed image;
step (4), an edge response fusion algorithm combining improved HED network edge detection and a watershed algorithm based on distance transformation is adopted to effectively segment the ore image;
and (5) calculating the granularity characteristic parameters of the ore by using the segmented image to obtain the granularity information of the ore.
Compared with the prior art, the conveyor belt ore granularity detection method based on the edge response fusion algorithm combines the improved HED edge detection method with the watershed algorithm based on distance transformation to obtain a more accurate conveyor belt ore image segmentation map, further obtains the granularity information of the ore, feeds the granularity information back to a control system of a crusher, and adjusts the crushing aperture of the crusher, the speed of the ore entering the crusher, the speed of a rotor and other factors which easily cause serious faults of the crusher. Accurate ore granularity information is obtained through calculating a segmentation graph, so that the fault rate of a crusher is greatly reduced, the production efficiency of a mine is improved, and the mining cost is saved.
Drawings
FIG. 1 is a flow chart of an implementation of the conveyor belt ore granularity detection method based on the edge response fusion algorithm.
Detailed Description
The embodiments of the present invention will be described in detail below with reference to the drawings and examples.
As shown in fig. 1, the invention relates to a conveyor belt ore granularity detection method based on an edge response fusion algorithm, which is based on a conveyor belt ore image granularity detection algorithm of a semantic segmentation technology, combines an improved HED network edge detection algorithm with a watershed algorithm based on distance transformation to obtain a segmentation image of an ore, and calls a partial function of OpenCV to obtain the granularity of the ore.
The method comprises the following specific steps:
and (1) installing an industrial camera on the conveyor belt, shooting an ore image on the conveyor belt in real time by using the industrial camera, transmitting the image into a background system, and generally randomly grabbing and shooting.
In order to obtain better segmentation effect, a conveyor belt ore photo with better quality needs to be taken firstly, the good photo can enable the pre-processed effect to be better, and the influence of external factors on the undesirable segmentation is reduced.
And (2) preprocessing the acquired image, including image graying, denoising and image binarization. Specifically, the method comprises the following steps:
a. the graying process is to adjust the gray value of each pixel point of the image to the average value of RGB components. That is, the average value of RGB of each pixel point of the ore image is obtained, so that the RGB values are equal, the image enhancement effect is achieved, and useful image information is enhanced.
b. Because the mine environment is severe, collected conveyor belt ore images are often full of noise, so that the segmentation effect in the later period is not ideal, and in order to solve the problem, the images are denoised by adopting a filtering method. The invention adopts a bilateral filtering method to carry out denoising, determines the value of an output pixel through the weighted combination of the pixel values of the neighborhood, and can achieve the best denoising effect while preserving the ore edge.
Bilateral filtering is a filtering method based on Gauss and has strong practicability. The specific method is that the brightness information of the image and the Gauss function are multiplied to optimize the filtering weight coefficient, and then the obtained coefficient and the image are subjected to convolution operation, so that the edge information of the image can be retained to the maximum extent. The denoising effect of bilateral filtering depends on the spatial distance and the pixel difference, and the weighted combination of the neighboring pixels can be represented by the following formula:
the weighting factor w (i, j, k, l) is determined by the product of the domain-defining kernel d (i, j, k, l) and the value domain kernel r (i, j, k, l).
σdAnd σrThe method is based on the standard deviation of a Gaussian function, is known as a spatial proximity factor and a brightness proximity factor, and can be used for controlling the performance of the bilateral filter. The weighted average value is used for replacing the gray value of the pixel, and the method can well represent that the gray value difference is smallThe pixel is suitable for different ore particles with similar colors. SigmadAnd σrIs a determining factor influencing the denoising effect, and therefore, how to take the appropriate value for the two is a key step of bilateral filtering. After a plurality of experiments, finally taking sigmad=30,σrThe filtering effect is optimal when the value is 0.05.
c. The image is binarized by using an integral image-based adaptive thresholding algorithm, the image is thresholded mainly by using an integral image method, and the finally obtained binarized image H (n) can be represented by the following formula:
the value of t in the formula is 15, and the mean value of the gray scale of the pixel point in the rectangular frame determines the threshold value of binarization, wherein the mean value of the gray scale (x, y) is determined by the following formula:
and (3) carrying out binarization image optimization and distance transformation on the preprocessed image. Specifically, the method comprises the following steps:
the method of the integral image is used for obtaining the binary image, and because the shape of the ore stone is extremely irregular, isolated points such as holes, burrs and the like still exist in the binary image, which can cause the result of segmentation to be inaccurate. The method comprises the following steps of optimizing a binary image by using a mathematical morphology method, carrying out expansion and corrosion operations on the image, and removing 'holes' and isolated noise points in the image, wherein the method mainly comprises the following steps:
the first step is as follows: and setting a structural element value for the image, and performing simple opening and closing operation on the image to reduce a small part of obvious holes and noise points.
The second step is that: and obtaining a complementary image of the original image by using an erosion reconstruction method.
The third step: and optimizing the complementary image, and then performing open operation to obtain a final optimized image.
Experiments prove that the binary image can be effectively optimized by using a mathematical morphology method, so that the edge part of the ore is smoother, and the expansion operation can effectively fill part of holes, so that the segmentation result is more accurate.
And then, performing distance transformation on the binary image, namely processing pixel points in the neighborhood of the image, and reducing over-segmentation.
And (4) adopting an edge response fusion algorithm combining improved HED network edge detection and a watershed algorithm based on distance transformation to perform real-time edge detection, and effectively segmenting the ore image. Specifically, the method comprises the following steps:
the improved HED network used by the invention is improved in a small part on the basis of the original network. Because the original HED network feature extraction layer is mainly based on the VGG network and includes a large number of pooling layers, the pooling layers can increase the receptive field while extracting image features, but reduce the size of the feature map, resulting in reduced resolution. Meanwhile, some edge information in the ore feature map is lost, which affects the segmentation accuracy.
The improved HED network deletes the third and fourth pooling layers, so that more network high-level image information with richer semantic features can be extracted, the extracted edges are finer, and the problems of over-segmentation and under-segmentation caused by ore adhesion can be well solved by combining a watershed algorithm based on distance transformation. Meanwhile, the deconvolution layer on the side edge is correspondingly modified; and replacing the remaining pooling layer with a hole convolution to increase the receptive field. The two-step improved operation not only can enable the extracted ore edge to be finer, but also reduces the workload of parameter design, obtains more ore image global information and provides an effective edge for subsequent edge fusion.
The expression for the hole convolution can be represented by:
n=k+(k-1)×(d-1)
wherein k is the size of the convolution kernel; p is the padding number in the convolution process; s is the convolution step length; d is the void convolution sampling rate; i is the size of the input feature map, and n and o are the size of a new convolution kernel and the size of an output feature map after the void convolution respectively;
the loss function of the improved HED network is:
wherein L isside(W, W) is the output cost of the edge, αmFor the weight of the loss function for each side,for the loss function of each side, W is the set of all parameters in the network, W is the weight parameter in the side output network, m is the current layer number, n is the total layer number, W(m)Is the network weight parameter output by the mth layer side.
The network optimizes each iteration by using a cross entropy function, and the expression is as follows:
(W,w,h)*=argmin(Lside(W,w)+Lfuse(W,w,h))
after a series of preprocessing, labeling the images by using labelme software, expanding the number of samples by a series of data enhancement operations such as cutting, reversing and zooming of the labeled images, finally dividing the data set into a training set and a testing set according to the proportion of 8:2, uniformly adjusting the size of pixels to 512x512, training 5000 times in the environment of 4 RTX2080 video cards with a programming language of Python and OpenCV of Python3.5, iterating once every 100 times, and setting the batch _ size to be 1. And reserving the optimal training weight, testing the test set and reserving the result.
Because conveyor belt ore is heavily stacked and ores of different sizes are stuck together, it is difficult to accurately segment the ore by one segmentation method alone, and there are severe over-segmentation and under-segmentation, so that the results obtained by the two methods are considered to be fused.
The improved HED network would fail to segment the stuck particles where the ore stacking is severe, while the watershed algorithm excels in segmenting the stuck particles, just as well as complementing them. The invention adopts a watershed algorithm of distance transformation to segment the adhered particles, and the essence of the method is that the distance transformation is matched with the watershed transformation, so that the direct watershed segmentation of the ore image is easy to generate over-segmentation. And the distance transformation is carried out on the obtained binary image, so that the over-segmentation can be effectively reduced. It mainly acts on the pixels in the neighborhood, and there are two general measurement methods, four neighborhood d(4)And eight neighborhoods d(8). The minimum proximity distance from a pixel background point p (j, k) to the foreground pixel point p (l, m) is calculated as follows:
d(4)(l,m)=min(|l-j|+|m-k|)
d(8)(l,m)=min{max(|l-j|),|m-k|}
l, m, j, k are coordinate values, d(4)(l, m) denotes the four neighbourhoods d(4)Minimum proximity distance, d, calculated by a metric method(8)(l, m) denotes the eight neighbourhood d(8)The minimum proximity distance calculated by the metric method.
And (3) applying a watershed algorithm to the distance-transformed binary image to obtain an ore segmentation image, fusing the ore segmentation image with the ore edge segmentation image obtained by adopting the improved HED network, wherein the fusion of the images is to make an AND algorithm for two results, and effectively merging the two segmentation images to obtain the final segmentation image. The watershed algorithm can effectively make up for the deficiency of the segmentation capability of the HED network at the adhesion part, and the HED network can make up for the characteristic that the design of watershed parameters is too complex.
And (5) analyzing and calculating the granularity characteristic parameters of the ore by using the segmented image to obtain the granularity information of the ore. Specifically, the method comprises the following steps:
and calling an OpenCV library, and calculating characteristic information such as the area, the perimeter, the total amount and the like of the ore particles. Finding all ore particle contours in the segmentation map by using a findContours algorithm; obtaining the minimum external rectangle of each ore outline by using a boundingRec algorithm; calculating the area of each contour by using a contourArea algorithm; the perimeter of each contour is calculated using the arcLength algorithm.
On this basis, can still further include:
and (6) transmitting the obtained ore granularity information to a crusher control system in real time, and adjusting parameters which can cause the crusher to break down, such as crushing caliber, feeding and discharging speed and the like according to the granularity.
In summary, the invention obtains the high-precision conveyor belt ore segmentation image through the fusion algorithm, and calculates the precise ore granularity according to the image, thereby providing important reference for adjusting the relevant parameters of the crusher, reducing the fault occurrence probability of the ore crusher, improving the production efficiency of the mine and saving the mining cost.
Claims (10)
1. A conveyor belt ore granularity detection method based on an edge response fusion algorithm is characterized by comprising the following steps:
step (1), using an industrial camera to shoot an ore image on a conveyor belt in real time;
step (2), preprocessing the collected image, including image graying, denoising and image binarization;
step (3), carrying out binarization image optimization and distance transformation on the preprocessed image;
step (4), an edge response fusion algorithm combining improved HED network edge detection and a watershed algorithm based on distance transformation is adopted to effectively segment the ore image;
and (5) calculating the granularity characteristic parameters of the ore by using the segmented image to obtain the granularity information of the ore.
2. The conveyor belt ore granularity detection method based on the edge response fusion algorithm as claimed in claim 1, wherein in the step (2), the graying of the image is to adjust the gray value of each pixel point of the image to the average value of RGB components; denoising, namely performing denoising processing on an image by using bilateral filtering, and determining the value of an output pixel through weighted combination of pixel values of neighborhoods; image binarization is realized by using an adaptive thresholding algorithm based on an integral image.
3. The conveyor belt ore granularity detection method based on the edge response fusion algorithm as claimed in claim 1 or 2, wherein in the step (3), the binarized image is optimized by using a morphological method, and the binarized image is subjected to distance transformation.
4. The conveyor belt ore granularity detection method based on the edge response fusion algorithm as claimed in claim 3, wherein the morphological method is to perform expansion and corrosion operations on the image to remove 'holes' and isolated noise points in the image, so that the image edge is smoother, the distance is converted into pixel points in the neighborhood of the processed image, and over-segmentation is reduced.
5. The conveyor belt ore granularity detection method based on the edge response fusion algorithm of claim 1, wherein the improved HED network is obtained by reducing two pooling layers on the basis of an HED original network, correspondingly modifying the deconvolution layers on the side edges, and replacing the remaining pooling layers with hole convolution to increase the receptive field.
6. The conveyor belt ore granularity detection method based on the edge response fusion algorithm is characterized in that the expression of the void convolution is as follows:
n=k+(k-1)×(d-1)
wherein k is the size of the convolution kernel; p is the padding number in the convolution process; s is the convolution step length; d is the void convolution sampling rate; i is the size of the input feature map, and n and o are the size of a new convolution kernel and the size of an output feature map after the void convolution respectively;
the loss function of the improved HED network is:
wherein L isside(W, W) is the output cost of the edge, αmFor the weight of the loss function for each side,for the loss function of each side, W is the set of all parameters in the network, W is the weight parameter in the side output network, m is the current layer number, n is the total layer number, W(m)Is the network weight parameter output by the mth layer side.
7. The conveyor belt ore granularity detection method based on the edge response fusion algorithm as claimed in claim 1, 5 or 6, wherein the distance-transformed binarized image is subjected to a watershed algorithm to obtain an ore segmentation image, and the ore segmentation image is fused with an ore edge segmentation image obtained by using an improved HED network to obtain a final segmentation image, wherein the fusion of the images is to perform an AND algorithm on two results.
8. The conveyor belt ore granularity detection method based on the edge response fusion algorithm of claim 7, wherein in the watershed algorithm, the minimum proximity distance from a pixel background point p (j, k) to a foreground pixel point p (l, m) is calculated as follows:
d(4)(l,m)=min(|l-j|+|m-k|)
d(8)(l,m)=min{max(|l-j|),|m-k|}
l, m, j, k are coordinate values, d(4)(l, m) denotes the four neighbourhoods d(4)Minimum proximity distance, d, calculated by a metric method(8)(l, m) denotes the eight neighbourhood d(8)The minimum proximity distance calculated by the metric method.
9. The conveyor belt ore granularity detection method based on the edge response fusion algorithm is characterized in that in the step (5), an OpenCV library is called, and all ore particle contours in a segmentation map are searched by using a findContours algorithm; obtaining the minimum external rectangle of each ore outline by using a boundingRec algorithm; calculating the area of each contour by using a contourArea algorithm; the perimeter of each contour is calculated using the arcLength algorithm.
10. The conveyor belt ore granularity detection method based on the edge response fusion algorithm as claimed in claim 1, wherein after the step (5), the obtained ore granularity information is transmitted to a crusher control system in real time, and parameters which can cause the crusher to malfunction are adjusted according to the granularity.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110603450.XA CN113240663B (en) | 2021-05-31 | 2021-05-31 | Conveyor belt ore granularity detection method based on edge response fusion algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110603450.XA CN113240663B (en) | 2021-05-31 | 2021-05-31 | Conveyor belt ore granularity detection method based on edge response fusion algorithm |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113240663A true CN113240663A (en) | 2021-08-10 |
CN113240663B CN113240663B (en) | 2023-03-14 |
Family
ID=77135903
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110603450.XA Active CN113240663B (en) | 2021-05-31 | 2021-05-31 | Conveyor belt ore granularity detection method based on edge response fusion algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113240663B (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114897789A (en) * | 2022-04-14 | 2022-08-12 | 安徽工业大学 | Sinter particle size detection method and system based on image segmentation |
CN115953776A (en) * | 2023-03-09 | 2023-04-11 | 聊城市检验检测中心 | Food detection system based on machine learning |
CN116823827A (en) * | 2023-08-29 | 2023-09-29 | 山东德信微粉有限公司 | Ore crushing effect evaluation method based on image processing |
CN116977827A (en) * | 2023-09-25 | 2023-10-31 | 中国检验认证集团山东有限公司 | Iron ore detection method and system based on artificial intelligence |
CN117274193A (en) * | 2023-09-21 | 2023-12-22 | 中铁十九局集团矿业投资有限公司 | Copper mine data analysis method and system based on artificial intelligence |
CN117994318A (en) * | 2024-04-07 | 2024-05-07 | 山东瑞福锂业有限公司 | Particle size detection method and system in ore conveying process |
CN117994318B (en) * | 2024-04-07 | 2024-06-07 | 山东瑞福锂业有限公司 | Particle size detection method and system in ore conveying process |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2016091016A1 (en) * | 2014-12-12 | 2016-06-16 | 山东大学 | Nucleus marker watershed transformation-based method for splitting adhered white blood cells |
EP3171297A1 (en) * | 2015-11-18 | 2017-05-24 | CentraleSupélec | Joint boundary detection image segmentation and object recognition using deep learning |
CN109214415A (en) * | 2018-07-19 | 2019-01-15 | 西安建筑科技大学 | A kind of ore reduction grinding effect evaluation method |
CN109785378A (en) * | 2018-03-21 | 2019-05-21 | 威海晶合数字矿山技术有限公司 | A kind of ore grain size online detection instrument based on atlas image algorithm analytical technology |
CN110570440A (en) * | 2019-07-19 | 2019-12-13 | 武汉珈和科技有限公司 | Image automatic segmentation method and device based on deep learning edge detection |
CN111047555A (en) * | 2019-11-13 | 2020-04-21 | 鞍钢集团矿业有限公司 | Ore image granularity detection algorithm based on image processing technology |
CN112365494A (en) * | 2020-11-30 | 2021-02-12 | 北京理工大学 | Ore material image segmentation method based on deep learning edge prediction |
-
2021
- 2021-05-31 CN CN202110603450.XA patent/CN113240663B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2016091016A1 (en) * | 2014-12-12 | 2016-06-16 | 山东大学 | Nucleus marker watershed transformation-based method for splitting adhered white blood cells |
EP3171297A1 (en) * | 2015-11-18 | 2017-05-24 | CentraleSupélec | Joint boundary detection image segmentation and object recognition using deep learning |
CN109785378A (en) * | 2018-03-21 | 2019-05-21 | 威海晶合数字矿山技术有限公司 | A kind of ore grain size online detection instrument based on atlas image algorithm analytical technology |
CN109214415A (en) * | 2018-07-19 | 2019-01-15 | 西安建筑科技大学 | A kind of ore reduction grinding effect evaluation method |
CN110570440A (en) * | 2019-07-19 | 2019-12-13 | 武汉珈和科技有限公司 | Image automatic segmentation method and device based on deep learning edge detection |
CN111047555A (en) * | 2019-11-13 | 2020-04-21 | 鞍钢集团矿业有限公司 | Ore image granularity detection algorithm based on image processing technology |
CN112365494A (en) * | 2020-11-30 | 2021-02-12 | 北京理工大学 | Ore material image segmentation method based on deep learning edge prediction |
Non-Patent Citations (5)
Title |
---|
LI YUAN等: "A Method of Ore Image Segmentation Based on Deep Learning", 《INTELLIGENT COMPUTING METHODOLOGIES (2018)》 * |
S.G.MKWELO等: "Watershed-based segmentation of rock scenes and proximity-based classification of watershed regions under uncontrolled lighting", 《SAIEE AFRICA RESEARCH JOURNAL 》 * |
董珂等: "基于改进分水岭变换的矿石图像分割算法", 《计算机工程与设计》 * |
蔡改贫等: "基于形态学优化处理的标记符分水岭矿石图像分割", 《科学技术与工程》 * |
顾清华等: "基于改进HED 网络模型的破碎矿石图像分割方法", 《激光与光电子学进展》 * |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114897789A (en) * | 2022-04-14 | 2022-08-12 | 安徽工业大学 | Sinter particle size detection method and system based on image segmentation |
CN115953776A (en) * | 2023-03-09 | 2023-04-11 | 聊城市检验检测中心 | Food detection system based on machine learning |
CN116823827A (en) * | 2023-08-29 | 2023-09-29 | 山东德信微粉有限公司 | Ore crushing effect evaluation method based on image processing |
CN116823827B (en) * | 2023-08-29 | 2023-11-10 | 山东德信微粉有限公司 | Ore crushing effect evaluation method based on image processing |
CN117274193A (en) * | 2023-09-21 | 2023-12-22 | 中铁十九局集团矿业投资有限公司 | Copper mine data analysis method and system based on artificial intelligence |
CN116977827A (en) * | 2023-09-25 | 2023-10-31 | 中国检验认证集团山东有限公司 | Iron ore detection method and system based on artificial intelligence |
CN117994318A (en) * | 2024-04-07 | 2024-05-07 | 山东瑞福锂业有限公司 | Particle size detection method and system in ore conveying process |
CN117994318B (en) * | 2024-04-07 | 2024-06-07 | 山东瑞福锂业有限公司 | Particle size detection method and system in ore conveying process |
Also Published As
Publication number | Publication date |
---|---|
CN113240663B (en) | 2023-03-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113240663B (en) | Conveyor belt ore granularity detection method based on edge response fusion algorithm | |
CN109961049B (en) | Cigarette brand identification method under complex scene | |
CN109472200B (en) | Intelligent sea surface rubbish detection method, system and storage medium | |
CN106780486A (en) | A kind of Surface Defects in Steel Plate image extraction method | |
CN109255326B (en) | Traffic scene smoke intelligent detection method based on multi-dimensional information feature fusion | |
CN107154044B (en) | Chinese food image segmentation method | |
WO2021118463A1 (en) | Defect detection in image space | |
CN101447027B (en) | Binaryzation method of magnetic code character area and application thereof | |
CN113256624A (en) | Continuous casting round billet defect detection method and device, electronic equipment and readable storage medium | |
Nejati et al. | License plate recognition based on edge histogram analysis and classifier ensemble | |
CN111754538A (en) | Threshold segmentation method for USB surface defect detection | |
CN111507391A (en) | Intelligent identification method for nonferrous metal broken materials based on machine vision | |
CN116452506A (en) | Underground gangue intelligent visual identification and separation method based on machine learning | |
CN115587966A (en) | Method and system for detecting whether parts are missing or not under condition of uneven illumination | |
Saifullah et al. | Fish Detection Using Morphological Approach Based-on K-Means Segmentation | |
CN113971681A (en) | Edge detection method for belt conveyor in complex environment | |
Lu et al. | Fusion-based color and depth image segmentation method for rocks on conveyor belt | |
CN108205678A (en) | A kind of nameplate Text region processing method containing speck interference | |
CN112785568A (en) | Magnetic shoe defect segmentation method based on deep learning | |
CN112288726A (en) | Method for detecting foreign matters on belt surface of underground belt conveyor | |
Zhu et al. | Optimization of image processing in video-based traffic monitoring | |
CN110751150A (en) | FPGA-based binary neural network license plate recognition method and system | |
Aung et al. | Study for license plate detection | |
CN114820718A (en) | Visual dynamic positioning and tracking algorithm | |
Graszka | Median mixture model for background–foreground segmentation in video sequences |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |