Rapid identification method for ground-air combined analysis of rock blasting effect

Abstract

The invention discloses a rapid identification method for ground-air combined analysis of a rock blasting effect, which is characterized by comprising the following steps of: collecting advantage information in a ground-air manner through an image processing technology of an unmanned aerial vehicle and an exploration vehicle, performing data comparison and calibration by taking the advantage information as a reference, and determining a muck pile area; establishing a visual three-dimensional explosion area by using a deep learning algorithm; then, the grade of rock in the muck pile area unit block is subjected to rapid field identification through the visual three-dimensional muck area; then estimating the rock grade of each unit block by using a distance power inverse ratio method according to the rock grade subjected to rapid field identification; and finally, estimating the rock grade of the whole muck pile area according to the rock grade of each unit block. By fusing the unmanned aerial vehicle and the all-terrain exploration unmanned vehicle, rapid acquisition and accurate analysis of the blasting muck pile particle size after open blasting are achieved, the efficiency and accuracy of blasting effect evaluation are remarkably improved, the problems of subjective errors and time consumption of a traditional method are solved, and technical support is provided for intelligent mine construction.

Description

Technical Field

[0001] The present invention relates to the technical field of open-pit mining applications, and particularly to a rapid identification method for analyzing the rock blasting effect by combining ground and air. Background Art

[0002] Traditional methods for evaluating the fragmentation effect of blasted rocks usually adopt the method of manual counting to detect the block size of the muck pile at the mine site. Although the advantage of this method is that it is simple to operate and does not require special training, it has some problems: for example, due to possible negligence and subjective judgment of on-site personnel, this method is prone to significant errors. Although it is convenient and fast to directly observe and record the rock accumulation form and block size distribution after blasting, it is greatly affected by subjectivity; even if multiple groups of personnel are used for recording and a multi-layer nested data loop comparison method is adopted, it is still not efficient enough when dealing with a large amount of data; at the same time, the manual method cannot accurately and quickly record the muck pile data, so it is impossible to evaluate the blasting effect in a timely manner, lacking effective guidance for subsequent blasting design and construction; there is also a large gap between the manual observation record and the actual block size distribution of the muck pile.

[0003] Therefore, to solve these problems, using image recognition technology and geographical location information to count the block size of the muck pile has become a feasible solution. Exploring the application of image recognition algorithms to the recognition of different muck pile images, establishing and optimizing the recognition model to achieve rapid, efficient, instant and on-site determination of the muck pile block size helps to build an intelligent mine and improve the accuracy and efficiency of blasting effect evaluation. Summary of the Invention

[0004] In order to solve the deficiencies of the above technologies, the present invention provides a rapid identification method for analyzing the rock blasting effect by combining ground and air.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is: a rapid identification method for analyzing the rock blasting effect by combining ground and air, which collects dominant information from the ground and air through the image processing technology of unmanned aerial vehicles and exploration vehicles, uses the dominant information as a reference for data comparison and calibration to determine the muck pile area; establishes a visible three-dimensional blasting area using a deep learning algorithm; then rapidly identifies the rock grade within the unit blocks of the muck pile area using the visible three-dimensional blasting area; then estimates the rock grade of each unit block using the distance power inverse ratio method based on the rock grade after rapid on-site identification; and finally estimates the rock grade of the entire muck pile area from the rock grades of each unit block.

[0006] Further, it specifically includes the following steps:

[0007] Step 1, set the parameters of the unmanned aerial vehicle to determine the target platform and scanning area;

[0008] Step 2, the all-terrain exploration vehicle divides the muck pile boundary;

[0009] Step 3, data preprocessing of the muck pile area;

[0010] Step 4, establish a visible three-dimensional blasting area;

[0011] Step 5, draw the grading curve and analyze the block size;

[0012] Step 6, evaluate the rock fragmentation effect.

[0013] Further, in Step 1, the flight parameters of the drone are as follows: the altitude is 4m to 8m, and the imaging degree ISO is not less than 150;

[0014] The settings of the flight record video file are as follows: the horizontal flight is set to an adaptive horizontal line flight attitude, and the angle setting feedback is not less than once per second. The imaging file is set to not less than 1080P / 30fps and not higher than 2k / 60fps. The drone automatically records the aircraft flight trajectory coordinate information, automatically marks the shadow surface and the ore-rock elevation information. The drone scans the muck pile based on the two-dimensional spatial quantity density calculation method, and takes not less than 1㎡ as the unit area, and marks the area with a quantity density lower than 2 / ㎡ as the non-muck pile area.

[0015] Further, the scanning area in Step 1 includes: the basic muck pile shape, the geographical location information of the muck pile, and the relative position information of the rocks; among them, the geographical location information of the muck pile is to output a file containing geographical location information and image information marked by the rock appearance. The relative position information of the rocks is the image information of different ore-rocks recorded as an image file with a marked frame line, and is merged with the file containing three-dimensional position information and recorded as the position file of different ore-rocks.

[0016] Further, Step 2 includes synchronizing the geographical location information in the drone scanning area. Based on the boundary line of the muck pile area after the drone scans, the traveling route of the exploration vehicle is set within the boundary line. The maximum walking range is set within a 1㎡ unit area. The exploration vehicle is set to continuously record the position information and output a.prg file. It interacts with the drone at a position not less than 0.5m and not higher than 1.5m outside the boundary line. The position of the exploration vehicle is set to be adaptive every 10s. The system compares the position of the drone frame line and the exploration vehicle every 5s, and the system returns the deviation value for the exploration vehicle to adjust. The exploration vehicle performs acoustic ranging on the ore-rock, and when the distance exceeds 1m, the position of the exploration vehicle at this time is marked as the non-muck pile area; the acoustic detection module is set to the standard rock exploration mode of once every 0.5s. The pitch angle of the on-vehicle stereo camera remains not less than 45° and not higher than 75° relative to the horizontal plane and is set to be adjusted not less than 2 times per second during movement. The on-vehicle imaging degree is set to ISO not less than 150.

[0017] Further, step 3 includes that the drone marks the muck pile area in the air with the quantity density as the evaluation criterion, and the exploration vehicle measures the distance by sound wave on the ground; the drone and the exploration vehicle divide the muck pile area, and the data comparison is carried out in the form of geographical location information. The 100% overlapping area will be directly determined as the muck pile area. The areas divided by the drone and the exploration vehicle are compared with a weight ratio of 60% and 40% to delete redundant data. The front and rear areas are combined into the final muck pile area, and the muck pile area information is saved in a file with geographical location information.

[0018] Further, step 4 includes:

[0019] Step 41, based on the driving route of the exploration vehicle and the geographical information segmentation coordinate points of the drone, plan the muck pile area, add the image files recorded by the drone in the divided muck pile area, and paste the image files with adaptive frame lines; according to the elevation information of the stereo camera of the exploration vehicle, adjust the height during 3D modeling with a single ore or rock as the unit.

[0020] Step 42, scan the overlapping ore or rock and compare it with the 3D position of the ore or rock in step 41 for model superposition. Based on the coordinate information of the single ore or rock of the drone, adjust the 3D model data and coordinate information of a single ore or rock with the elevation information and coordinate information of the ore or rock measured by the exploration vehicle.

[0021] Further, step 5 includes:

[0022] Step 51, set the stope muck pile area as several target unit areas, and calculate the unit area size; then calculate the radius r of the inscribed circle of the unit block with the center of the unit area of the block to be identified as the center, and determine the area within the inscribed circle as the detection range; determine the relative size of the block according to (r1 + r2) / 2, where r1 is the farthest distance between the edge of each block finally measured and falling within the detection range and the detection range, and r2 is the closest distance between the edge of each block finally measured and falling within the detection range and the detection range.

[0023] Step 52, based on step 51, take the largest block and the smallest block as the upper and lower limits (H), record the total number of rock blocks as (W), take H / W as the scale of the grading curve, count the number of blocks in each proportion interval and draw the grading curve.

[0024] Step 53, according to the obtained grading curve, estimate different particle sizes by using the following formula:

[0025]

[0026] In formula (1) and formula (2), λ iIt represents the weight coefficient, which is the inverse of the k-th power of the distance from the center of the target ore-rock; gi represents the mass of the i-th ore-rock; di represents the distance from the i-th ore-rock to the center of the block size to be estimated; g represents the block size of the unit to be identified, that is, each unit block size is equal to the weighted average of each ore-rock within its influence range.

[0027] Further, in step 6, the average block size of the entire blasting area is obtained by recalculating the distance power inverse method for each unit ore-rock as follows:

[0028] Xc = (∫∫xdσ) / A Formula (3);

[0029] Yc = (∫∫ydσ) / A Formula (4);

[0030] A = ∫∫dσ Formula (5);

[0031] In Formula (3), Formula (4) and Formula (5), A is the irregular detection area; Xc is the abscissa of the centroid; Yc is the ordinate of the centroid; x is the abscissa of the center of gravity; y is the ordinate of the center of gravity;

[0032] Among them, if the detection area is an irregular graph, the centroid is taken as the center, and a point on the graph is selected as the origin. The x-axis coordinate is obtained in Formula (3), and the y-axis coordinate is obtained in Formula (4). Among them, A can be solved by Formula (5) as the area of the irregular graph. After determining the detection center, measure the distance from the center of each target block size to the boundary of the detection area;

[0033]

[0034] In Formula (6) and Formula (7), γ i is the weight coefficient, which is the inverse of the k-th power of the distance from the center of the unit block to the center of the area to be identified; gj is the ore-rock size of the j-th unit block; Di is the distance from the i-th unit ore-rock to the detection center to be estimated; G is the particle size of the ore in the detection area to be identified, that is, the particle size of the iron ore in each blasting area is equal to the weighted average of the particle sizes of each unit ore within its influence range.

[0035] Further, the digital images of ores with different particle sizes are numbered in groups, and numbered in sequence from low block size to high block size at intervals of 1%; the blasted pile area is set as several target unit areas, and the size of the unit area and the radius of the inscribed circle of the calibration area are obtained by combining the following formulas for calculation:

[0036]

[0037] nx is the number of ore-rocks in the x direction within the unit block; ny is the number of ore-rocks in the y direction within the unit block; N is the total number of ore-rocks in the blasting area; T is the number of target unit blocks;

[0038] Substitute data N and T into formula (8) to obtain n x ×n y , that is, the number of ore and rock in each unit area;

[0039] Let nx = ny to obtain nx and ny, and substitute nx and ny into formula (9) and formula (10) respectively to obtain the unit area size and the radius R of the inscribed circle as follows:

[0040]

[0041] Lx is the unit area; Ly is the width of the unit area; a is the spacing between ore and rock; b is the area distance; R is the radius of the inscribed circle.

[0042] The present invention discloses a rapid recognition method for analyzing the rock blasting effect in a ground-air combination manner. By using image processing and deep learning algorithms, it realizes the rapid recognition of the muck pile and the accurate estimation of the iron ore grade, reduces the subjective error and detection time, improves the recognition accuracy, and provides technical support for the construction of intelligent mines. Brief Description of the Drawings

[0043] Figure 1 is a schematic flow chart of the rapid recognition method of the present invention.

[0044] Figure 2 is a schematic structural view of the exploration vehicle of the present invention.

[0045] Figure 3 is a schematic structural view of the unmanned aerial vehicle of the present invention.

[0046] In the figure: 1, stereo camera imaging processing module; 2, imaging processing module; 3, acoustic detector; 4, motion sensor; 5, central processor; 6, exploration vehicle battery; 7, lidar; 8, coordinate positioning module; 9, central processor; 10, camera; 11, unmanned aerial vehicle battery; 12, propeller. Detailed Description of the Invention

[0047] The present invention will be further described in detail below with reference to the drawings and specific embodiments.

[0048] First, as Figure 2 and 3The ground-air combined analysis of the rapid identification method of rock blasting effect of the present invention relies on an exploration vehicle that can exchange data with a drone. The exploration vehicle is used for ground exploration of the blast pile morphology. The exploration vehicle includes a stereo camera imaging processing module 1 for photographing the blast pile morphology and the ore rock stacking situation at the ground pitch and elevation angles; the imaging processing module 2 is used to verify and compare the data collected by the stereo camera imaging processing module 1 and output a non-redundant image file; the acoustic wave detector 3 is used to accurately measure the relative position relationship between the exploration vehicle and the blast pile; the motion sensor 4 is used to detect the travel state of the exploration vehicle including direction, angle, speed, etc.; the central processing unit 5 is used to process interactive files and control the overall operation; the exploration vehicle battery 6 is used to power the exploration vehicle electrical components; the laser radar 7 is used to indicate the exploration vehicle's travel route; the drone includes a coordinate positioning module 8, a central processing unit 9, a camera 10, a drone battery 11 and a propeller 12. Among them, the coordinate positioning module 8 is responsible for providing accurate geographical location information for the drone, ensuring that the drone can accurately locate the location of the blast pile when scanning in the air, and matching the collected data with the ground exploration vehicle for data comparison and calibration. As the core component of the UAV, the central processor 9 is responsible for processing all input data, including positioning information, image data, etc. It is also responsible for coordinating various functions of the UAV, such as flight control, data processing, data transmission, etc. The camera 10 is a device used by the UAV for ground aerial photography. Its main function is to collect image information of the morphology of the blast pile and the stacking of ore rocks. These image data will be used for subsequent image processing and data analysis to quickly identify the particle size and block size distribution of the blast pile. The UAV battery 11 provides the required power for the UAV to ensure that the UAV can continue to work during the flight mission and data processing. The propeller 12 is responsible for providing the UAV with flight power, enabling it to perform stable flight and scanning operations in the air. The design and performance of the propeller directly affect the stability of the UAV flight and the accuracy of the collected data.

[0049] Based on this, the present invention discloses a rapid identification method for analyzing rock blasting effects by ground-to-air combined analysis, which collects advantageous information from the ground and air through image processing technology of unmanned aerial vehicles and exploration vehicles, and uses the advantageous information as a reference to perform data comparison and calibration to determine the blasting area; a visual three-dimensional blasting area is established using a deep learning algorithm; the visible three-dimensional blasting area is then used to perform rapid on-site identification of the rock grade within the unit block in the blasting area; the distance power inverse method is then used to estimate the rock grade of each unit block based on the rock grade after rapid on-site identification; finally, the rock grade of the entire blasting area is estimated from the rock grade of each unit block.

[0050] The specific steps include:

[0051] Step 1: Set the parameters of the drone to determine the target platform and scanning area; the flight parameter height of the drone is 4m to 8m, and the imaging ISO is not less than 150;

[0052] The settings of the flight record video file are as follows: for horizontal flight, the flight attitude and angle are adaptively set to provide feedback no less than once per second. The imaging file is set to be no less than 1080P / 30fps and no higher than 2k / 60fps. The drone automatically records the flight trajectory coordinate information of the aircraft, automatically marks the shadow surface and the elevation information of the ore and rock. The drone scans the muck pile based on the two-dimensional spatial quantity density calculation method, and takes no less than 1㎡ as the unit area, and marks the area with a quantity density lower than 2 / ㎡ as the non-muck pile area.

[0053] The scanning area includes: the basic muck pile shape, the geographical location information of the muck pile, and the relative position information of the rocks; among them, the geographical location information of the muck pile is to output a file containing geographical location information and image information marked by the appearance of the rocks. The relative position information of the rocks is the image information of different ores and rocks recorded as an image file with identification frame lines, and is merged with the file containing three-dimensional position information and recorded as the position file of different ores and rocks.

[0054] Step 2, the all-terrain exploration vehicle demarcates the boundary of the muck pile; Step 2 includes synchronizing the geographical location information in the scanning area of the drone. Based on the boundary line of the muck pile area after the drone scans, the travel route of the exploration vehicle is determined. The maximum walking range is set within a unit area of 1㎡. The exploration vehicle is set to continuously record the position information and output a.prg file. It conducts coordinate interaction with the drone at a position not less than 0.5m and not higher than 1.5m outside the boundary line. The position of the exploration vehicle is set to be adaptively adjusted at 10s intervals. The system compares the position of the drone's frame line with that of the exploration vehicle once every 5s. The system returns the deviation value for the exploration vehicle to adjust. The exploration vehicle conducts acoustic ranging on the ore and rock, and marks the position of the exploration vehicle at this time as the non-muck pile area when the distance exceeds 1m; the acoustic detection module is set to the standard rock exploration mode of once every 0.5s. The pitch angle of the on-vehicle stereo camera is maintained at no less than 45° and no higher than 75° relative to the horizontal plane and is set to be adjusted no less than 2 times per second during movement. The on-vehicle imaging degree is set to ISO not less than 150.

[0055] Step 3, preprocess the data of the muck pile area; Step 3 includes the drone marking the muck pile area in the air with the quantity density as the evaluation standard, and the exploration vehicle using acoustic ranging as the evaluation standard on the ground; the drone and the exploration vehicle demarcate the muck pile area and conduct data comparison in the form of geographical location information. The 100% overlapping area will be directly determined as the muck pile area. The areas demarcated by the drone and the exploration vehicle are compared with a weight ratio of 60% and 40% to delete redundant data. The front and rear areas are combined into the final muck pile area, and the muck pile area information is saved in a file with geographical information.

[0056] Step 4, establish a visible three-dimensional muck pile area; Step 4 includes:

[0057] Step 41: Based on the driving route of the exploration vehicle and the segmentation coordinate points of the UAV geographic information, plan the muck pile area. Add the image files recorded by the UAV to the divided muck pile area and paste the image files with adaptive bounding boxes. According to the elevation information of the stereo camera of the exploration vehicle, adjust the height during 3D modeling with a single ore or rock as a unit.

[0058] Step 42: Scan the overlapping ore or rock and compare it with the 3D positions of the ore or rock in Step 41 for model superposition. Using the coordinate information of a single ore or rock of the UAV as the main body, adjust the 3D model data and coordinate information of a single ore or rock according to the elevation information and coordinate information of the ore or rock measured by the exploration vehicle.

[0059] Step 5: Draw the grading curve and perform fragmentation analysis. Step 5 includes:

[0060] Step 51: Set the blasting area of the stope as several target unit areas and calculate the size of the unit area. Then, taking the center of the unit area of the block size to be identified as the center, calculate the radius r of the inscribed circle of the unit block, and determine the area within the inscribed circle as the detection range. Determine the relative size of the block according to (r1 + r2) / 2, where r1 is the maximum distance from the edge of each block finally measured falling within the detection range to the detection range, and r2 is the minimum distance from the edge of each block finally measured falling within the detection range to the detection range.

[0061] Step 52: Based on Step 51, take the maximum block size and the minimum block size as the upper and lower limits (H), record the total number of rock blocks as (W), take H / W as the scale of the grading curve, count the number of blocks in each proportion interval and draw the grading curve.

[0062] Step 53: According to the obtained grading curve, estimate different particle sizes using the following formula:

[0063]

[0064] In Formula (1) and Formula (2), λ i represents the weight coefficient, that is, the inverse ratio of the kth power of the distance from the center of the target ore or rock; gi represents the mass of the ith ore or rock; di represents the distance from the ith ore or rock to the center of the block to be estimated; g represents the block size of the unit to be identified, that is, each unit block size is equal to the weighted average of each ore or rock within its influence range.

[0065] Step 6: Evaluate the rock fragmentation effect. · In Step 6, the average block size of the entire blasting area is obtained by re-calculating each unit ore or rock using the distance power inverse method as follows:

[0066] Xc = (∫∫xdσ) / A Formula (3);

[0067] Yc = (∫∫ydσ) / A Formula (4);

[0068] A = ∫∫dσ Equation (5);

[0069] In Equations (3), (4) and (5), A is the irregular detection area; Xc is the abscissa of the centroid; Yc is the ordinate of the centroid; x is the abscissa of the center of gravity; y is the ordinate of the center of gravity;

[0070] Among them, if the detection area is an irregular figure, the centroid is taken as the center, and a point on the figure is selected as the origin. The x-axis coordinate is obtained in Equation (3), and the y-axis coordinate is obtained in Equation (4). Among them, A can be solved by Equation (5) as the area of the irregular figure. After determining the detection center, measure the distance from the center of each target block size to the boundary of the detection area.

[0071]

[0072] In Equations (6) and (7), γ i is the weight coefficient, which is the inverse of the k-th power of the distance from the center of the unit block to the center of the area to be recognized, that is, the distance; gj is the size of the ore and rock of the j-th unit block; Di is the distance from the i-th unit ore and rock to the center of the detection to be estimated; G is the particle size of the ore in the detection area to be recognized, that is, the particle size of the iron ore in each blasting area is equal to the weighted average of the particle sizes of each unit ore within its influence range;

[0073] Furthermore, the digital images of ores with different particle sizes are numbered in groups, and numbered sequentially from low block size to high block size at intervals of 1%; the muck pile area is set as several target unit areas, and the size of the unit area and the radius of the inscribed circle of the calibration area are obtained by combining and calculating using the following formula:

[0074]

[0075] nx is the number of ore and rock in the x direction within the unit block; ny is the number of ore and rock in the y direction within the unit block; N is the total number of ore and rock in the blasting area; T is the number of target unit blocks;

[0076] Substitute the data N and T into Equation (8) to obtain n x ×n y , that is, the number of ore and rock in each unit area;

[0077] Let nx = ny, and obtain nx and ny. Substitute nx and ny into Equations (9) and (10) respectively to obtain the unit area size and the radius R of the inscribed circle as follows:

[0078]

[0079] Lx is the unit area; Ly is the width of the unit area; a is the spacing of ore and rock; b is the area distance; R is the radius of the inscribed circle.

[0080] In a specific embodiment of open-pit marble mining, an SWDE1521 intelligent drill is used for drilling, with an average hole depth of 16m, an overdepth of 1.5m, and a hole diameter of 140mm. After blasting, the blasted area is scanned and data is processed. First, blasting design and on-site blasting are carried out. Then, the drone and the exploration vehicle enter the muck pile area for operation. After the drone scans the muck pile and shares the data with the exploration vehicle, the exploration vehicle calculates the travel route based on the scanned blasted area range of the drone. The exploration vehicle measures the real-time distance from the exploration vehicle to the muck pile according to the acoustic detector equipped on itself. At a position not less than 0.5m and not higher than 1.5m outside the boundary line, a stereo camera module is used to collect images of the muck pile. The imaging processing module converts the image information into a three-dimensional model for establishing a visible three-dimensional muck pile, and uses its own motion sensor to record the geographical information position, and identifies and corrects the shared data to divide the fine muck pile spatial geographical information. Then, information interaction and comparison are carried out to establish a blasted area model and draw a grading curve. Finally, the blasting effect is evaluated based on the evaluation criteria.

[0081] Referring to Table 1 of the crushing effect grade evaluation in this field, the parameter index is 25 / ㎡ and the average block size is 360mm, which meets the grade of 20 / ㎡ < N, 400mm < L < 100mm in the grade evaluation. It solves the disadvantages of complex measurement process, large calculation amount, and long time consumption of the on-site muck pile. The single blasted area identification time only takes 10 minutes. Compared with manual counting, the accuracy rate of the grading curve can reach 95%, realizing the rapid on-site identification of the particle size distribution of the ore and rock in the open-pit blasting muck pile, laying a foundation for building an intelligent mine.

[0082] Table 1 Crushing effect grade evaluation

[0083]

[0084] To sum up, the present invention provides a method for collecting the particle size of the muck pile and quickly analyzing the blasting effect of a drone and an all-terrain exploration unmanned vehicle based on a stereo camera and a ground acoustic sensor after open-pit blasting. Geographical location information is obtained through aerial scanning by the drone, the coordinate information of the ore and rock is marked and the muck pile range is divided. The stereo camera scans to obtain the position vector information between each ore and rock. The all-terrain exploration unmanned vehicle scans and records the muck pile information again from the horizontal angle relative to the muck pile, judges the lithology distribution of the muck pile using the principle of acoustic flaw detection, and combines the data collected from the ground and the air for data comparison and data verification, so as to achieve the purpose of accurately identifying the ore and rock and reducing the missed detection of the ore and rock.

[0085] Compared with the prior art, the advantages of the present invention are:

[0086] 1) The present invention is based on image processing technology to quickly identify the muck pile. By comparing the geographical location information of the unmanned aerial vehicle and the exploration vehicle, redundant information is deleted, and the information advantages of the unmanned aerial vehicle and the exploration vehicle are combined. Taking the advantageous information as a reference for data comparison and calibration, the modeling texture map is made to better conform to the digital images of the samples collected separately from the on-site muck pile. Then, a deep learning algorithm is used to establish an iron ore grade identification model. Subsequently, the iron ore grade in the blast hole of each unit block in the blast area is quickly identified on-site using this identification model. Then, the iron ore grade of each unit block is estimated by using the distance power inverse method based on this grade. Finally, the iron ore grade of the entire blast area is estimated from the iron ore grades of each unit block.

[0087] 2) The present invention solves the problems such as subjective error and long time consumption in the determination of the particle size and block size distribution of the on-site muck pile, realizes the quick on-site identification and estimation of the muck pile block size, and lays a foundation for building an intelligent mine.

[0088] 3) The present invention conducts an immediate on-site determination of the muck pile block size distribution. This process does not require manual chemical analysis, and the identification time for a single blast area is only 10 minutes. Compared with manual visual inspection, the absolute error of the muck pile block size distribution is 0.01, that is, the accuracy rate of the model for identifying the muck pile block size can reach 90%.

[0089] The above embodiments are not limitations to the present invention, and the present invention is not limited to the above examples. Any changes, modifications, additions, or substitutions made by those skilled in the art within the scope of the technical solution of the present invention also fall within the protection scope of the present invention.

Claims

1. A rapid identification method for analyzing rock blasting effects by ground-air combined analysis, characterized in that: The image processing technology of drones and exploration vehicles is used to collect advantageous information from the ground and air, and the advantageous information is used as a reference for data comparison and calibration to determine the explosion area; a visual three-dimensional explosion area is established using a deep learning algorithm; the visual three-dimensional explosion area is then used to quickly identify the rock grade within the unit block in the explosion area on site; the inverse power method of distance is then used to estimate the rock grade of each unit block based on the rock grade after rapid on-site identification; finally, the rock grade of the entire explosion area is estimated from the rock grade of each unit block.

2. The rapid identification method for analyzing rock blasting effects by ground-air combined analysis according to claim 1 is characterized in that: The specific steps include: Step 1: Set the parameters of the drone to determine the target platform and scanning area; Step 2: The all-terrain exploration vehicle demarcates the blast pile boundary; Step 3: preprocessing of data in the explosion area; Step 4, establishing a visible three-dimensional explosion area; Step 5: gradation curve drawing and block size analysis; Step 6: rock crushing effect evaluation.

3. The rapid identification method for analyzing rock blasting effects by ground-air combined analysis according to claim 2 is characterized in that: In step 1, the flight parameter height of the drone is 4m to 8m, and the imaging ISO is not less than 150; The settings of the flight recording video file are: horizontal flight is adaptive horizontal line flight attitude and angle setting feedback is not less than once per second, the imaging file is set to be not less than 1080P / 30fps and not higher than 2k / 60fps, the UAV automatically records the aircraft flight trajectory coordinate information, automatically marks the shadow surface and ore rock elevation information, the UAV scans the explosive pile based on the two-dimensional space number density calculation method, and takes not less than 1㎡ as the unit area, and marks the number density below 2 / ㎡ as non-explosion pile area.

4. The rapid identification method for analyzing rock blasting effects by ground-air combined analysis according to claim 2 is characterized in that: The scanning area in step 1 includes: basic blast pile morphology, blast pile geographical location information, and rock relative position information; wherein, the blast pile geographical location information is a file containing geographical location information and image information output by marking the rock appearance, and the rock relative position information is the image information of different mineral rocks recorded as an image file containing identification frame lines, and is merged with the file containing three-dimensional position information to be recorded as a position file of different mineral rocks.

5. The rapid identification method for analyzing rock blasting effects by ground-to-air combined analysis according to claim 2 is characterized by: The step 2 includes synchronizing the geographic location information in the area scanned by the drone, according to which the boundary line of the explosive pile area scanned by the drone is the route of the exploration vehicle, and the unit area of ​​1 square meter of quantity density is the maximum walking range, setting the exploration vehicle to continuously record the position information and output a .prg file, interacting with the drone at a position not less than 0.5m and not more than 1.5m outside the boundary line, setting the exploration vehicle to adapt to the position at intervals of 10s, comparing the position of the drone border line and the exploration vehicle 5s / time in the system, and adjusting the deviation value returned by the system by the exploration vehicle, and the exploration vehicle performing acoustic ranging of the ore and rock, and marking the position of the exploration vehicle at this time as a non-explosive pile area when the distance exceeds 1m; setting the acoustic detection module to the standard 0.5s / time rock exploration mode, maintaining the pitch angle of the on-board stereo camera relative to the horizontal plane at not less than 45° and not more than 75° and setting it to be adjusted at least twice per second with movement, and setting the on-board imaging degree to ISO not less than 150.

6. The rapid identification method for analyzing rock blasting effects by ground-to-air combined analysis according to claim 2 is characterized by: The step 3 includes the UAV marking the explosion pile area in the air based on the number density as the evaluation standard, and the exploration vehicle using the acoustic ranging as the evaluation standard on the ground; the UAV and the exploration vehicle divide the explosion pile area and compare the data in the form of geographic location information. The 100% overlapping area will be directly determined as the explosion pile area, and the area divided by the UAV and the area divided by the exploration vehicle will be compared with each other at a weight ratio of 60% and 40% to delete redundant data, and the front and back areas will be combined into the final explosion pile area, and the explosion area information will be saved in a file with geographic information location.

7. The rapid identification method for analyzing rock blasting effects by ground-to-air combined analysis according to claim 2 is characterized by: The step 4 comprises: Step 41, based on the exploration vehicle's route and the segmented coordinate points of the UAV's geographic information, the blast pile area is planned, the image file recorded by the UAV is added to the divided blast pile area, and the image file is mapped with adaptive frame lines; according to the elevation information of the exploration vehicle's stereo camera, the height adjustment is performed when three-dimensional modeling is performed with a single ore rock as a unit; Step 42, scan the overlapping mineral rocks and compare the three-dimensional positions of the mineral rocks in step 41 to perform model superposition, taking the coordinate information of the single mineral rock of the drone as the main body, and adjusting the three-dimensional model data and coordinate information of the single mineral rock with the elevation information and coordinate information of the mineral rock measured by the exploration vehicle.

8. The rapid identification method for analyzing rock blasting effects by ground-air combined analysis according to claim 2 is characterized in that: The step 5 comprises: Step 51, set the blasting area of ​​the stope as a number of target unit areas, and calculate the size of the unit area; then calculate the radius r of the inscribed circle of the unit block with the center of the unit area of ​​the block to be identified as the center, and determine the area within the inscribed circle as the detection range; determine the relative size of the block according to (r1+r2) / 2, where r1 is the farthest distance between the edge of each block and the detection range, and r2 is the shortest distance between the edge of each block and the detection range; Step 52, based on step 51, the maximum block size and the minimum block size are used as the upper and lower limits (H), the total number of rock blocks is recorded as (W), H / W is used as the scale of the grading curve, the number of blocks in each ratio interval is counted and the grading curve is drawn; Step 53, based on the obtained grading curve, use the following formula to estimate different particle sizes: In formula (1) and formula (2), λ i It represents the weight coefficient, which is the inverse of the kth power of the distance from the center of the target ore rock; gi represents the mass of the ith ore rock; di represents the distance of the ith ore rock from the center of the blockiness to be estimated; g represents the blockiness of the unit to be identified, that is, the blockiness of each unit is equal to the weighted average of all the ores within its influence range.

9. The rapid identification method for analyzing rock blasting effects by ground-to-air combined analysis according to claim 2 is characterized by: In step 6, the average fragmentation of the entire blasting area is obtained by recalculating each unit ore rock using the inverse power method of distance as follows: Xc=(∫∫xdσ) / A Formula (3); Yc=(∫∫ydσ) / A Formula (4); A=∫∫dσ formula (5); In formula (3), formula (4) and formula (5), A is the irregular detection area; Xc is the abscissa of the centroid; Yc is the ordinate of the centroid; x is the abscissa of the center of gravity; y is the ordinate of the center of gravity; If the detection area is an irregular shape, the centroid is the center. A point in the shape is selected as the origin. The x-axis coordinate is obtained in formula (3), and the y-axis coordinate is obtained in formula (4). A can be solved as the area of ​​the irregular shape by formula (5). After determining the detection center, measure the distance from the center of each target block to the boundary of the detection area. In formula (6) and formula (7), γ i is the weight coefficient, which is the inverse ratio of the distance from the center of the unit block to the center of the area to be identified, i.e. the distance to the kth power; gj is the ore size of the jth unit block; Di is the distance from the ith unit ore to the center of the detection to be estimated; G is the ore particle size of the detection area to be identified, i.e. the iron ore particle size of each blasting area is equal to the weighted average of the ore particle sizes of each unit within its influence range.

10. The rapid identification method for analyzing rock blasting effects by ground-to-air combined analysis according to claim 9 is characterized by: Digital images of ores with different particle sizes are numbered in groups and numbered in sequence from low to high particle size at intervals of 1%. The blast area is set as several target unit areas, and the unit area size and the radius of the inscribed circle of the calibration area are calculated using the following formula: nx is the number of mineral rocks in the x direction within the unit block; ny is the number of mineral rocks in the y direction within the unit block; N is the total number of mineral rocks in the blasting area; T is the number of target unit blocks; Substituting data N and T into formula (8), we get n x ×n y , that is, the number of mineral rocks in each unit area; Let nx = ny, and get nx and ny. Substitute nx and ny into formula (9) and formula (10) respectively to get the unit area size and the radius R of the inscribed circle as follows: Lx is the unit area; Ly is the unit area width; a is the distance between ore and rock; b is the area distance; R is the radius of the inscribed circle.

Images