CN118196168A - Method and device for calculating length of dry beach of tailing pond, storage medium and product - Google Patents

Abstract

The invention discloses a method, a device, a storage medium and a product for calculating the length of a dry beach of a tailing pond, and relates to the technical field of operation and safety monitoring of the tailing pond; performing edge contour detection extraction on the dry beach area by using an image processing technology; fitting the identified front dam ore drawing edge contour into a straight line by using a straight line fitting method; obtaining the maximum value, the minimum value and the average value of the pixel distances according to the pixel distances from each pixel point on the dry beach water area dividing line to the ore drawing edge in front of the dam fitted into a straight line; and converting the maximum value, the minimum value and the average value of the pixel distance into actual distances according to the resolution ratio and the scale of the satellite image, and obtaining the maximum length, the minimum length and the average length of the dry beach of the tailing pond. The invention reduces the influence of external environmental factors to the greatest extent, can replace manual inspection, and improves the measurement efficiency and accuracy.

Description

A method, device, storage medium and product for calculating the dry beach length of a tailings pond

Technical Field

The present invention relates to the technical field of tailings pond operation and safety monitoring, and in particular to a method, device, storage medium and product for calculating the dry beach length of a tailings pond.

Background technique

Tailings ponds are indispensable facilities for mining production. During the mining process, the ore dressing plant will extract useful substances from the original rough ore according to certain processes, and a large amount of tailings waste will be generated at the same time. Tailings are a by-product of ore processing, and the tailings pond is a place to store tailings waste in mining production.

When processing tailings, the tailings will be mixed with water to form tailings slurry, which will then be transported to the tailings pond through the ore discharge pipeline for discharge. Because the discharged tailings slurry contains a large amount of water, after sedimentation, a tidal flat area and a water area will be formed in the tailings pond. The tidal flat area is also called the dry beach area. The dry beach line refers to the boundary between the tidal flat area and the water area, and the dry beach length refers to the straight-line distance from the ore discharge edge of the reservoir area to the boundary between the tidal flat and the water area.

The length of the dry beach of the tailings pond is an important indicator for measuring the safe operation of the tailings pond. If the indicator exceeds the safety threshold, it may cause major safety accidents such as water overflowing in the tailings pond and dam collapse, which will pose a threat to the safety of people’s lives and property. Therefore, it is very important to accurately obtain the length of the dry beach of the tailings pond.

Among the existing dry beach length measurement technologies, there are methods that use manually placed markers to estimate the distance; there are also methods that use cameras in the reservoir area to estimate the dry beach length, etc. For example: A Chinese patent (application number 201710491350.6) discloses a method for measuring the dry beach length of a tailings pond using machine vision, in which three markers are set up on the dry beach of the tailings, and the three markers are in a straight line, and the straight line is perpendicular to the dam body; the distances between the three markers and the distances from each marker to the dam body are measured respectively; a digital image is taken using a color high-definition digital camera, and the positions of the three markers in the digital image are extracted; the digital image is divided into a dry beach area and a water area; a straight line is fitted according to the position coordinates of the three markers in the digital image, and the intersection ratio of the marker position points and the boundary intersection points on the fitted straight line is calculated; the dry beach distance is calculated using the property of the invariant cross ratio. For another example: A Chinese patent (application number 201911380099.1) discloses a dry beach measurement method based on photogrammetry and laser ranging, including using a remote automatic photography system to calibrate an industrial camera and regularly take photos of the dry beach; based on the dry beach photos, using a waterline automatic recognition algorithm to accurately identify the position of the waterline; using a precise control servo system to adjust the angle of the laser ranging device so that the laser ranging device is aligned with the identified precise waterline position; the laser ranging device automatically measures the distance from the waterline to the laser lens, and obtains the dry beach length through geometric conversion. For another example: A Chinese patent (application number 202310264510.9) discloses a method for calculating the dry beach length of a tailings pond based on deep learning, including using a trained deep learning image recognition model to identify the dam and water edge area in the image, and calculate the image pixel distance between the dam and the water edge of the water accumulation area at the measurement section to obtain the distance pixel ratio at the section; according to the distance pixel ratio at the section, calculate the distance from the dam body to the water edge of the water accumulation area at the measurement section; calculate the difference between the beach length and the beach top water level difference; finally, calculate the dry beach length of the tailings pond based on the beach length and the beach top water level difference.

However, the above methods have the disadvantages of many calculation steps and complicated implementation process. In addition, it is still necessary to go to the site to install cameras, insert scales, set up measuring instruments, etc. These steps are easily affected by the external environment, require a lot of manual operations, and are greatly affected by interference factors.

Therefore, in order to minimize the impact of external environmental factors and replace manual inspections to improve measurement efficiency and accuracy, it is urgent to provide a new method for calculating the dry beach length of tailings ponds.

Summary of the invention

The purpose of the present invention is to provide a method, device, storage medium and product for calculating the dry beach length of a tailings pond, which can reduce the influence of external environmental factors to the greatest extent, and at the same time can replace manual inspections and improve measurement efficiency and accuracy.

To achieve the above object, the present invention provides the following solutions:

A method for calculating the dry beach length of a tailings pond, the method comprising:

The trained deep learning model Mask R-CNN is used to identify the entire dry beach area of the tailings pond in the satellite image; the dry beach area is a polygonal area surrounded by the dam front ore edge, the reservoir boundary and the dry beach water dividing line;

Image processing technology is used to detect and extract the edge contour of the dry beach area; the extracted edge contours include: the edge contour of the ore discharge in front of the dam, the dry beach water area dividing line and the reservoir area boundary;

The identified edge contour of the dam is fitted into a straight line using the straight line fitting method;

According to the pixel distance from each pixel point on the dry beach water dividing line to the dam front mining edge fitted as a straight line, the maximum, minimum and average values of the pixel distance are obtained;

According to the resolution and scale of the satellite image, the maximum, minimum and average values of the pixel distance are converted into actual distances to obtain the maximum, minimum and average lengths of the dry beach of the tailings pond.

Optionally, the trained deep learning model Mask R-CNN includes: a backbone network consisting of a residual neural network and a feature pyramid network, a region proposal network, and a region of interest alignment network.

Optionally, the training process of the deep learning model MaskR-CNN is:

Use labelme to mark the dry beach area of the tailings pond in the satellite image;

After labeling is completed, labelme generates a COCO format dataset;

Train the deep learning model Mask R-CNN based on the COCO format dataset.

Optionally, the edge contour detection and extraction method is a Canny algorithm.

Optionally, the image processing technique is a least squares method.

A computer device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method.

A computer-readable storage medium stores a computer program, which implements the steps of the method when executed by a processor.

A computer program product comprises a computer program, which, when executed by a processor, implements the steps of the method.

According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

The present invention provides a method, device, storage medium and product for calculating the dry beach length of a tailings pond, which use satellite images and deep learning models to efficiently identify the dry beach area and dry beach length, are less affected by external environmental factors, effectively reduce the impact of external factors such as weather and environment, and improve monitoring efficiency and accuracy; the present invention does not require on-site manual inspections and drone photography, and compared with the on-site video monitoring method for calculating the dry beach length, the present invention can reduce the measurement error caused by the camera shooting angle; the present invention uses satellite images, and does not need to set up cameras on site to obtain images, which is easy to operate and maintain in the later stage, and is simpler and more convenient than the existing technology; the calculation steps are simple and efficient; and it has a high engineering application value. In summary, the test effect of the present invention is good and the solution is feasible.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

FIG1 is a schematic flow chart of a method for calculating the dry beach length of a tailings pond provided by the present invention;

FIG2 is a schematic diagram of the Mask-RCNN network structure used in the present invention;

FIG3 is a schematic diagram showing the effect of the present invention on identifying dry beach areas in satellite images;

FIG4 is a schematic diagram showing the effect of grayscale processing of a recognized image according to the present invention;

FIG5 is a schematic diagram showing the effect of performing contour extraction on a grayscale processed image according to the present invention;

FIG6 is a schematic diagram showing the effect of retaining the contour of the ore drawing edge and the dry beach line of the present invention;

7 is a schematic diagram showing the effect of the present invention on fitting a straight line to the ore-drawing edge, establishing a coordinate axis, and calculating the straight-line distance from the dry beach line to the ore-drawing edge.

FIG. 8 is a diagram showing the internal structure of a computer device.

Detailed ways

The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

The purpose of the present invention is to provide a method, device, storage medium and product for calculating the dry beach length of a tailings pond, which can reduce the influence of external environmental factors to the greatest extent, and at the same time can replace manual inspections and improve measurement efficiency and accuracy.

In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

As shown in FIG1 , the present invention provides a method for calculating the dry beach length of a tailings pond, the method comprising:

S101, using the trained deep learning model MaskR-CNN, identifies the entire dry beach area of the tailings pond in the satellite image; the dry beach area is a polygonal area surrounded by the dam front ore edge, the reservoir boundary and the dry beach water dividing line;

As shown in Figure 2, Mask R-CNN is an instance segmentation algorithm that can be used for "target detection", "target instance segmentation", and "target key point detection". It is an extension of Faster R-CNN that outputs high-quality instance segmentation masks while effectively detecting targets. It adds a branch for predicting segmentation masks based on Faster R-CNN. Mask R-CNN has achieved excellent results in tasks such as instance segmentation, target detection, and human key point detection. The mainstream frameworks of Mask R-CNN are PyTorch and Tensorflow, and common Mask R-CNN dataset formats include COCO, PascalVOC, and so on. The deep learning framework used in the present invention is based on Tensorflow and Keras, and the dataset format is COCO.

The trained deep learning model MaskR-CNN includes: a backbone network consisting of a residual neural network and a feature pyramid network, a region proposal network, and a region of interest alignment network.

The ResNet (residual neural network) and FPN (feature pyramid network) in the backbone network extract features from the input image and generate feature maps; the region proposal network (RPN) generates anchor frames to extract areas that may contain target information and generates proposal frames of different sizes; ROIAlign (region of interest alignment network) adjusts the proposal frames of different sizes in the previous step to a uniform size to facilitate subsequent detection frame positioning and classification; the uniform-sized proposal frames are input into the mask branch and the detection frame branch respectively to generate the target mask and detection frame and their classification. The deep learning framework used in the present invention is Tensorflow, which is an open source machine learning tool launched by Google in 2015. The programming interface supports Python and C++ and is one of the most popular deep learning frameworks. The training set format of the data set used in the present invention is COCO; the programming language is Python.

The satellite images required for training the deep learning model Mask R-CNN can be obtained using map software such as Google Earth, Amap, and Baidu Map. The images are three-channel RGB images.

The training process of the deep learning model Mask R-CNN is:

S110, labeling the dry beach area of the tailings pond in the satellite image using the labelme identification tool;

The specific steps for labeling are as follows: open the labelme software, select the picture to be labeled, click the CreatePolygons button in the project bar on the left to start drawing polygons, and draw the dry beach area in the picture along the boundary; place the mouse on the first point, and click the second point to automatically close the boundary between the two points. After drawing a circle around the dry beach area boundary into a closed figure, the category box will automatically pop up. After customizing the category, click Save. After saving, the corresponding json data will be automatically generated; repeat the above steps until all data sets are labeled.

S111, after labeling, labelme generates a COCO format dataset;

Finally, open the json conversion file that comes with the labelme software, find and run json_to_dataset.py to generate the COCO format and complete the production of the dataset.

S112, train the deep learning model MaskR-CNN based on the COCO format dataset.

The COCO format dataset is generated by the image labeling tool labelme. The collected satellite images are labeled one by one using the labelme software, and the dry beach areas in the images are marked. The labeled images are used as the image training set. The number of images can be around 300. The more images there are, the more accurate the model prediction will be.

The labeled image training set is input into the Mask RCNN network for training. The number of training steps for each epoch is set to 1000, the number of verification steps for each epoch is set to 100, the threshold for filtering RPN proposals is set to 0.7, the learning rate is set to 0.001, the weight decay is set to 0.0001, and the training is performed for 100 rounds.

If you find that the accuracy is not as expected or the training is overfitting, you can adjust the number of training times, adjust parameter settings, and expand the data set to improve it.

After the training is completed, the training model weights will be generated. In the computer, the training model weights are represented as a file. For subsequent predictions, the network will use the training weight file as the core to predict the image to be predicted.

As shown in Figure 3, the image to be predicted is input into the Mask RCNN network. The network will automatically identify the dry beach area of the tailings pond in the image and generate a recognition result. The dry beach area will be covered by a red mask.

S102, using image processing technology to detect and extract edge contours of the dry beach area; the extracted edge contours include: the edge contour of the ore discharge in front of the dam, the dry beach water area dividing line and the reservoir area boundary;

For the satellite image after identification, grayscale processing and canny algorithm are used to extract the contour of the dry beach area; the specific steps are as follows: the first step is to perform grayscale processing and import cv2 and numpy libraries; first load the identified image (the dry beach area identified by the Mask R-CNN model in this invention is a red area by default, so the image processing method given later is to extract the contour of the red area; if the image recognition result shows that the area is set to other colors, define the corresponding HSV color space range), convert the image to HSV color space; secondly, define the upper and lower limits of the red range, create a mask, and set the remaining areas to black; finally, extract the red identification area of the dry beach through bitwise AND operation, and the rest of the background becomes black. The second step is to perform edge contour extraction processing, using cv2 and numpy libraries; first load the image and convert it to HSV color space; secondly define the red range in HSV space; create a mask layer, set the red part to white, and the rest to black; use double threshold tracking to determine the boundary, and extract the edge contour of the area, which is the contour of the dry beach area of the tailings pond. Finally, the contour line of the dry beach area is displayed as white, and the rest is a black background.

As shown in FIG. 4 , the image predicted in the above 7 is subjected to image grayscale processing, the identified red mask area of the dry beach is retained, and the remaining background parts are subjected to grayscale processing and are no longer retained but turned into black.

When performing grayscale processing, cv2 and numpy libraries are used. First, load the image.

Convert the image to HSV color space; define the upper and lower bounds of the red range, create a mask, and set the rest of the area to black; finally, extract the red identification area through a bitwise AND operation, and turn the rest of the background to black.

As shown in FIG. 5 , the dry beach area extracted in 9 above is further processed to extract its edge contour.

When performing edge contour extraction, cv2 and numpy libraries are used. First, load the image and convert it to HSV color space; second, define the red range in HSV space; create a mask layer, set the red part to white and the rest to black; finally, use the canny edge detection algorithm to extract the regional edge contour, which is the dry beach area contour of the tailings pond. Finally, the dry beach area contour line is displayed as white, and the rest is a black background.

As shown in Figure 6, the surrounding redundant contour lines are removed, leaving only the mining edge and the dry beach line. There are many methods, such as python programming, matlab programming, image processor, etc.

S103, using a straight line fitting method to fit the identified edge contour of the dam front ore drawing into a straight line;

The least squares method is used to fit a straight line to the line of the mining edge. The specific steps are as follows: first, the pixel coordinates of the mining edge line contour are extracted. Because the final contour of the mining edge is displayed in white, the function np.where(threshold == 255) is defined to return a tuple containing the coordinates of the white points. Then, the row index array and column index array are obtained through indexing operations to obtain the coordinate information; then, the np.polyfit function is used to fit the obtained data points to a straight line. Finally, the np.poly1d function is used to create a polynomial object to fit a straight line.

As shown in Figure 7, the least square method is used to perform a straight line fitting on the ore edge in front of the dam.

S104, obtaining the maximum value, minimum value and average value of the pixel distances from each pixel point on the dry beach water area dividing line to the dam front ore placement edge fitted as a straight line;

Traverse each pixel point on the dry beach water segmentation curve and calculate the distance from each pixel point to the straight line after fitting.

S105, converting the maximum value, minimum value and average value of the pixel distance into actual distance according to the resolution and scale of the satellite image, and obtaining the maximum length, minimum length and average length of the dry beach of the tailings pond.

The image used is a 16-level satellite image with an image resolution of 72dpi, 118 pixels, an on-image distance of 4.16cm, an actual distance of 500m, and a scale of 12000:1. Based on the scale information, the pixel distance of the dry beach in the image is converted into the actual distance to obtain the actual distance of the dry beach length.

In summary, the method of the present invention has the following characteristics: 1. Good adaptability and little interference from external environmental factors. 2. Convenient for monitoring and evaluating data. 3. Convenient for later operation and maintenance. 4. Simple operation, suitable for enterprise technicians to operate.

In one embodiment, a computer device is provided, which may be a database, and its internal structure diagram may be shown in FIG8. The computer device includes a processor, a memory, an input/output interface (Input/Output, referred to as I/O) and a communication interface. The processor, the memory and the input/output interface are connected via a system bus, and the communication interface is connected to the system bus via the input/output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store pending transactions. The input/output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a data processing method is implemented.

In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

It should be noted that the object information (including but not limited to object device information, object personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the object or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions.

Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

This article uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only used to help understand the method and core ideas of the present invention. At the same time, for those skilled in the art, according to the ideas of the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims (8)

1. A method for calculating the length of a dry beach of a tailings pond, comprising the steps of:

Identifying the whole dry beach area of the tailing pond in the satellite image by using a trained deep learning model Mask R-CNN; the dry beach area is a polygonal area surrounded by a front ore drawing edge of the dam, a reservoir area boundary and a dry beach water area dividing line;

Performing edge contour detection extraction on the dry beach area by using an image processing technology; the extracted edge profile includes: a front ore drawing edge contour of the dam, a dry beach water area dividing line and a reservoir area boundary;

fitting the identified front dam ore drawing edge contour into a straight line by using a straight line fitting method;

Obtaining the maximum value, the minimum value and the average value of the pixel distances according to the pixel distances from each pixel point on the dry beach water area dividing line to the ore drawing edge in front of the dam fitted into a straight line;

And converting the maximum value, the minimum value and the average value of the pixel distance into actual distances according to the resolution ratio and the scale of the satellite image, and obtaining the maximum length, the minimum length and the average length of the dry beach of the tailing pond.

2. The method for calculating the length of a dry beach of a tailings pond according to claim 1, wherein the trained deep learning model MaskR-CNN comprises: a backbone network formed by the residual neural network and the characteristic pyramid network, a region suggestion network and a region of interest alignment network.

3. The method for calculating the length of a dry beach of a tailings pond according to claim 1, wherein the training process of the deep learning model Mask R-CNN is as follows:

Marking a dry beach area of a tailing pond in a satellite image by utilizing labelme identification tools;

Generating a COCO format dataset by labelme after marking is finished;

Training a deep learning model Mask R-CNN according to the COCO format data set.

4. The method for calculating the length of a dry beach of a tailings pond according to claim 1, wherein the method for detecting and extracting the edge profile is a canny algorithm.

5. The method for calculating the length of a dry beach of a tailings pond according to claim 1, wherein the image processing technique is a least square method.

6. A computer apparatus, comprising: a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that the processor executes the computer program to implement the steps of the method of any one of claims 1-5.

7. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, carries out the steps of the method according to any one of claims 1-5.

8. A computer program product comprising a computer program, characterized in that the computer program/instructions, when executed by a processor, implement the steps of the method of any of claims 1-5.