A real-time rock type determination method during drilling process

Abstract

The present invention provides a method for determining rock type in real time during the drilling process, belonging to the field of mining drilling. The method comprises: installing a stroboscopic camera in the cone drill operation room to monitor the pointer changes of a rotary tachometer, a rotary ammeter, and a feed pressure gauge; transmitting the obtained instrument image to a terminal; performing grayscale and binarization processing on the instrument image; setting a polar coordinate system based on the instrument panel scale and pointer; establishing a matching relationship between the scale and the polar angle θ; recording the monitored data in a rock-ore matching relationship dataset; training a neural network model; and ultimately achieving real-time determination of the rock type within the blasthole using the instrument image. The present invention can further achieve real-time rock type determination during the cone drill drilling process, saving labor costs, improving the efficiency of blasthole rock type determination, and thus enhancing the economic benefits of the mine.

Description

Technical Field

[0001] The present invention relates to the technical field of large-scale open-pit mine rock drilling, and in particular to a method for real-time rock type judgment during a drilling process. Background Art

[0002] At mining blasting sites, traditional methods for measuring rock properties within blastholes typically use in-hole cameras or mineral powder sampling and testing. These methods are technically complex and costly, and identification technology is limited to post-blasthole completion, making real-time identification of ore rock properties difficult. Real-time drilling parameters such as rotary current, rotary speed, and feed pressure during drilling are closely aligned with ore rock properties and have garnered considerable attention and application. However, obtaining these parameters currently requires intelligent retrofitting of traditional rotary drills, relying on control system modifications and sensor installation, resulting in high maintenance costs and significant added expense. Summary of the Invention

[0003] The present invention provides a method for determining rock types in real time during a drilling process, comprising a stroboscopic camera and a data transmission terminal, and is characterized in that it comprises the following steps:

[0004] Step 1: Prepare rock specimens, conduct uniaxial compressive strength tests, and establish a database of mine rock strength and rock types;

[0005] Step 2: Install a 4G or Wi-Fi-based stroboscopic camera in front of the rotary drill instrument panel to periodically capture the rotary drill instrument panel during operation and generate images;

[0006] Step 3: First, the acquired image is transmitted to the data transmission terminal;

[0007] Step 4: Generate a grayscale image of the dashboard image in the data transmission terminal;

[0008] Step 5: Use the Ostu method to calculate the image binarization threshold, where T is the segmentation threshold for foreground and background. The number of foreground points in the image is w0, and the average grayscale is u0; the number of background points in the image is w1, and the average grayscale is u1. The total average grayscale of the image is u, and the variance g of the foreground and background images is: When the variance g is the largest, the grayscale T at this time is the optimal threshold;

[0009] Step 6: Binarize the dashboard image data according to the threshold calculated in step 5;

[0010] Step 7: Set the direction parallel to the 0 scale as the positive direction of the polar coordinate axis, select the starting point of the pointer as the pole, and establish the polar coordinate system;

[0011] Step 8: Set the scale corresponding to each polar angle. The minimum unit of polar angle selection is the angle when the instrument's minimum range changes, and establish the corresponding relationship between polar angle and scale value.

[0012] Step 9: Use the existing Sobel operator edge detection method to identify the pointer position of the binary image in step 5. The operator contains two sets of 3*3 matrices, one for horizontal and one for vertical. By performing a planar convolution with the image, the horizontal and vertical brightness difference approximations can be obtained respectively.

[0013] Step 10: Based on the polar coordinate system established in step 7, the pointer position identified in step 9 is converted into a scale value using the corresponding relationship in step 8;

[0014] Step 11: Obtain the scale value through the image and output the real-time drilling rotation speed curve, rotation current curve and feed pressure curve;

[0015] Step 12: Establish a ore-rock matching relationship data set, which records rock type, uniaxial compressive strength, rotary speed, rotary current, and feed pressure;

[0016] Step 13: Based on the data set established in step 12, a neural network training model is constructed with the rotation speed curve, the rotation current curve, and the feed pressure curve as the input layer, and the rock type and uniaxial compressive strength as the output layer. The model is trained to obtain a trained neural network model.

[0017] Step 14: In subsequent drilling, repeat steps 2 to 11, and input the real-time collected rotation speed curve, rotation current curve, and feed pressure curve into the trained neural network model to achieve real-time rock type judgment during the drilling process.

[0018] Preferably, the rotary drill instrument panel includes a rotary tachometer, a rotary ammeter and a feed pressure gauge.

[0019] Preferably, in step 7, the polar coordinate system is established with the center of the instrument panel as the pole and the positive direction of the x-axis as the polar axis direction, and an appropriate polar diameter ρ is selected to draw a circle, and the relationship between the polar angle θ and the corresponding scale and the position of the intersection of the pointer and the circle are obtained according to the instrument panel; the instrument panel image is binarized to establish the polar coordinate system.

[0020] The beneficial effects of the present invention are:

[0021] The present invention utilizes a stroboscopic camera mounted on a rotary drill to periodically photograph the instrument panel during operation, transmits the photographed images to a data terminal for grayscale and binarization processing, further analyzes and processes the data by establishing a polar coordinate system, establishes a mineral-rock matching relationship for the processed data, and trains a neural network training model, thereby achieving real-time judgment of the rock properties in the blasthole, facilitating the continuity and convenience of the rotary drill operation, and thus improving operating efficiency.

[0022] The method of the present invention can effectively improve work efficiency, significantly reduce manual labor intensity, and rapidly enhance digitalization and intelligence. The present invention can ultimately achieve the goal of solving the problem of real-time and accurate judgment of the lithology of the blasthole during the drilling process in the mine blasting area. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 This is a flow chart of the method for real-time rock type determination during the drilling process of the present invention.

[0024] Figure 2 This is a schematic diagram of the installation of the stroboscopic camera of the present invention. DETAILED DESCRIPTION

[0025] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described 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 making creative efforts should fall within the scope of protection of the present invention.

[0026] Example 1

[0027] uncle Figure 1 and Figure 2 As shown, in an iron ore mining plant, when a rotary drill is drilling a blasthole, a strobe camera 2 is installed in front of the instrument panel 1 to take regular snapshots at 10 Hz. The drilling time is 30 minutes, and 1,000 images are sent back by the on-site camera for processing.

[0028] The present invention provides a method for real-time rock type determination during a drilling process, comprising the following steps:

[0029] Step 1: Select 100 rock specimens from the mine and prepare them. Perform uniaxial compressive strength tests on all rock specimens and establish a database of rock strength and rock types based on the test results.

[0030] Table 1 Rock type and strength matching table

[0031]

[0032] Step 2: Install a 4G or Wi-Fi-based stroboscopic camera 2 in front of the rotary drill's rotation speed meter, rotary current meter, and feed pressure meter, respectively, to periodically capture and generate images of the rotary drill's instrument panel during operation.

[0033] Step 3: First, wirelessly transmit the acquired series of images to the data terminal 3;

[0034] Step 4: Generate grayscale images of all images in the terminal;

[0035] Step 5: Use the Ostu method to calculate the image binarization threshold, where T is the segmentation threshold for foreground and background. The number of foreground points in the image is w0, and the average grayscale is u0; the number of background points in the image is w1, and the average grayscale is u1. The total average grayscale of the image is u, and the variance g of the foreground and background images is:

[0036] u=w0×u0+w1×u1

[0037] g=w0×(u0-u) 2 +w1×(u1-u) 2

[0038] Combining the two formulas, we can get: When the variance g is the largest, the grayscale T is the optimal threshold;

[0039] Step 6: Binarize the instrument panel image data according to the threshold value calculated in step 5 to distinguish the instrument panel and the pointer;

[0040] Step 7: Establish a polar coordinate system with the center of the instrument panel as the pole and the 0 scale direction as the positive direction of the polar coordinate axis;

[0041] Step 8: Set the scale corresponding to each polar angle. The minimum unit of polar angle selection is the angle when the instrument's minimum range changes, and establish the corresponding relationship between polar angle and scale value.

[0042] Step 9: Use the existing Sobel operator edge detection method to identify the pointer position of the binary image in step 5. The operator contains two sets of 3*3 matrices, one for horizontal and one for vertical. By performing a planar convolution with the image, the horizontal and vertical brightness difference approximations can be obtained respectively.

[0043] Step 10: Based on the polar coordinate system established in step 7, the pointer position identified in step 9 is converted into a scale value using the corresponding relationship in step 8;

[0044] Step 11: Obtain the scale value through the image and output the real-time drilling rotation speed curve, rotation current curve and feed pressure curve;

[0045] Step 12: Establish a ore-rock matching relationship data set, which records rock type, uniaxial compressive strength, rotary speed, rotary current, and feed pressure;

[0046] Table 2 Matching relationship record table

[0047]

[0048] Step 13: Based on the data set established in step 12, with the rotation speed curve, rotation current curve, and feed pressure curve as the input layer, and the rock type and uniaxial compressive strength as the output layer, a neural network training model is built. The training is performed for more than 100 times to obtain a trained neural network model.

[0049] Step 14: During subsequent drilling, the real-time collected rotation speed curve, rotation current curve, and feed pressure curve are input into the trained neural network model to achieve real-time rock type judgment during the drilling process.

[0050] Example 2

[0051] At an iron ore mining plant, when a rotary drill is drilling a blasthole, a stroboscopic camera is installed in front of the dashboard to take regular snapshots at 10Hz. The drilling time is 60 minutes, and 2,000 pictures are sent back by the on-site camera for processing.

[0052] The present invention provides a method for real-time rock type determination during a drilling process, comprising the following steps:

[0053] Step 1: Select 200 rock specimens from the mine and prepare them. Perform uniaxial compressive strength tests on all rock specimens and establish a database of rock strength and rock types based on the test results.

[0054] Table 1 Rock type and strength matching table

[0055]

[0056] Step 2: Install 4G or Wi-Fi-based stroboscopic cameras in front of the rotary drill's tachometer, rotary ammeter, and feed pressure gauge, and periodically capture and generate images of the rotary drill's instrument panel during operation.

[0057] Step 3: First, wirelessly transmit the acquired series of images to the data terminal;

[0058] Step 4: Generate grayscale images of all images in the terminal;

[0059] Step 5: Use the Ostu method to calculate the image binarization threshold, where T is the segmentation threshold for foreground and background. The number of foreground points in the image is w0, and the average grayscale is u0; the number of background points in the image is w1, and the average grayscale is u1. The total average grayscale of the image is u, and the variance g of the foreground and background images is:

[0060] u=w0×u0+w1×u1

[0061] g=w0×(u0-u) 2 +w1×(u1-u) 2

[0062] Combining the two formulas, we can get: When the variance g is the largest, the grayscale T is the optimal threshold;

[0063] Step 6: Binarize the instrument panel image data according to the threshold value calculated in step 5 to distinguish the instrument panel and the pointer;

[0064] Step 7: Establish a polar coordinate system with the center of the instrument panel as the pole and the 0 scale direction as the positive direction of the polar coordinate axis;

[0065] Step 8: Set the scale corresponding to each polar angle. The minimum unit of polar angle selection is the angle when the instrument's minimum range changes, and establish the corresponding relationship between polar angle and scale value.

[0066] Step 9: Use the existing Sobel operator edge detection method to identify the pointer position of the binary image in step 5. The operator contains two sets of 3*3 matrices, one for horizontal and one for vertical. By performing a planar convolution with the image, the horizontal and vertical brightness difference approximations can be obtained respectively.

[0067] Step 10: Based on the polar coordinate system established in step 7, the pointer position identified in step 9 is converted into a scale value using the corresponding relationship in step 8;

[0068] Step 11: Obtain the scale value through the image and output the real-time drilling rotation speed curve, rotation current curve and feed pressure curve;

[0069] Step 12: Establish a ore-rock matching relationship data set, which records rock type, uniaxial compressive strength, rotary speed, rotary current, and feed pressure;

[0070] Table 2 Matching relationship record table

[0071]

[0072] Step 13: Based on the data set established in step 12, a neural network training model is built with the rotation speed curve, rotation current curve, and feed pressure curve as the input layer, and the rock type and uniaxial compressive strength as the output layer. The training is performed for more than 200 times to obtain a trained neural network model.

[0073] Step 14: During subsequent drilling, the real-time collected rotation speed curve, rotation current curve, and feed pressure curve are input into the trained neural network model to achieve real-time rock type judgment during the drilling process.

[0074] The present invention installs a stroboscopic camera in the cone drill operation room to monitor the pointer changes of the rotary tachometer, rotary ammeter and feed pressure gauge, transmits the obtained instrument image to the terminal, performs grayscale and binarization processing on the instrument panel, sets a polar coordinate system according to the instrument panel scale and pointer, establishes a matching relationship between the scale and the polar angle θ, records the monitored data in a mineral-rock matching relationship data set, trains a neural network model, and ultimately realizes real-time judgment of the rock properties in the blasthole through the instrument image.

Claims

1. A method for real-time rock type identification during drilling, comprising a stroboscopic camera and a data transmission terminal, characterized in that: Step 1: Prepare rock specimens, conduct uniaxial compressive strength tests, and establish a database of mine rock strength and rock types; Step 2: Install a 4G or Wi-Fi-based stroboscopic camera in front of the rotary drill instrument panel to periodically capture the rotary drill instrument panel during operation and generate images; Step 3: First, the acquired image is transmitted to the data transmission terminal; Step 4: Generate a grayscale image of the dashboard image in the data transmission terminal; Step 5: Use the Ostu method to calculate the image binarization threshold, where T is the segmentation threshold for foreground and background. The number of foreground points in the image is w0, and the average grayscale is u0; the number of background points in the image is w1, and the average grayscale is u1. The total average grayscale of the image is u, and the variance g of the foreground and background images is: When the variance g is the largest, the grayscale T at this time is the optimal threshold; Step 6: Binarize the dashboard image data according to the threshold calculated in step 5; Step 7: Set the direction parallel to the 0 scale as the positive direction of the polar coordinate axis, select the starting point of the pointer as the pole, and establish the polar coordinate system; In step 7, the polar coordinate system is established with the center of the instrument panel as the pole and the positive direction of the x-axis as the polar axis. The polar coordinate system is established, an appropriate polar diameter ρ is selected to draw a circle, and the relationship between the polar angle θ and the corresponding scale and the position of the intersection of the pointer and the circle are obtained according to the instrument panel; the instrument panel image is binarized to establish the polar coordinate system; Step 8: Set the scale corresponding to each polar angle. The minimum unit of polar angle selection is the angle when the instrument's minimum range changes, and establish the corresponding relationship between polar angle and scale value. Step 9: Use the existing Sobel operator edge detection method to identify the pointer position of the binary image in step 5. The operator contains two sets of 3*3 matrices, one for horizontal and one for vertical. By performing a planar convolution with the image, the horizontal and vertical brightness difference approximations can be obtained respectively. Step 10: Based on the polar coordinate system established in step 7, the pointer position identified in step 9 is converted into a scale value using the corresponding relationship in step 8; Step 11: Obtain the scale value through the image and output the real-time drilling rotation speed curve, rotation current curve and feed pressure curve; Step 12: Establish a ore-rock matching relationship data set, which records rock type, uniaxial compressive strength, rotary speed, rotary current, and feed pressure; Step 13: Based on the data set established in step 12, a neural network training model is constructed with the rotation speed curve, the rotation current curve, and the feed pressure curve as the input layer, and the rock type and uniaxial compressive strength as the output layer. The model is trained to obtain a trained neural network model. Step 14: In subsequent drilling, repeat steps 2 to 11, and input the real-time collected rotation speed curve, rotation current curve, and feed pressure curve into the trained neural network model to achieve real-time rock type judgment during the drilling process.

2. The method for real-time rock type identification during drilling according to claim 1, characterized in that: The rotary drill instrument panel includes a rotary tachometer, a rotary ammeter and a feed pressure gauge.

Images