Ore sorting system and method based on multi-light-path compensation

Abstract

The invention discloses an ore sorting system and method based on multi-light-path compensation, relates to the technical field of ore analysis, and solves the technical problems that in the prior art, when ore is recognized, the recognition result is inaccurate and ore sorting is not facilitated due to interference of factors such as the surface condition of the ore and light change. The method comprises the following steps: acquiring real-time environment data and mining area data of an ore to-be-sorted area; setting light path parameters of multiple light paths according to the mining area data of the to-be-sorted ore area; adjusting the parameters of the multiple light paths based on the real-time environment data to obtain adjusted parameters; dividing the mining area into a plurality of detection areas according to the mining area data; detecting the plurality of detection areas according to the adjustment parameters to obtain detection data; sorting the ores according to the detection data; light path parameters can be adjusted according to a real-time environment, interference of environmental factors on identification is reduced, and the accuracy of a sorting result is improved.

Description

Technical Field

[0001] The present invention belongs to the field of ore analysis and relates to an ore sorting technology based on multi-optical path compensation, in particular to an ore sorting system and method based on multi-optical path compensation. Background Art

[0002] Ore sorting is a key link in the mining and mineral processing fields. Its core lies in utilizing the differences in the physical and chemical properties of minerals to efficiently separate useful minerals from gangue. Multi-optical path compensation technology is an advanced ore sorting method. Its core lies in the synergy of multiple optical paths, combined with intelligent algorithms, to achieve accurate identification of the physical and chemical properties of ores and efficient sorting. The multi-optical path system uses multiple light sources or sensors to illuminate the ore surface from different angles, which can effectively eliminate the shadows and blind spots caused by the irregular shape, uneven surface or uneven reflection of the ore in a single light path. The multi-optical path system can offset the effects of ambient light changes, equipment vibration or dust interference in real time by automatically adjusting the sensitivity of the light source or sensor. It is suitable for complex working conditions on mining sites and avoids sorting interruptions caused by environmental fluctuations.

[0003] The prior art (invention patent application with publication number: CN118926129A) discloses a method for identifying ore, which includes: obtaining multiple images of a stone to be tested at different angles; determining the ore image information in each of the images; if the ore image information is not detected in the multiple images, determining that the stone to be tested is waste rock; if the ore image information is detected in at least one of the images, determining the classification of the stone to be tested based on the ore image information of each of the images; the prior art accurately identifies the type of ore by collecting multiple images at different angles, which can reduce the occurrence of identification omissions; however, when identifying the ore, the prior art will be interfered by factors such as the surface condition of the ore and changes in light, resulting in inaccurate identification results, which is not conducive to ore sorting.

[0004] The present invention provides an ore sorting system and method based on multi-optical path compensation to solve the above technical problems. Summary of the Invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes an ore sorting system and method based on multi-optical path compensation, which is used to solve the technical problem that the prior art is affected by factors such as the ore surface condition and light changes when identifying ore, resulting in inaccurate identification results, which is not conducive to ore sorting.

[0006] To achieve the above-mentioned object, a first aspect of the present invention provides an ore sorting system based on multi-optical path compensation, comprising: an ore detection module, and a data acquisition module and a terminal sorting module connected thereto;

[0007] Data acquisition module: used to obtain real-time environmental data of the ore sorting area and mining area data;

[0008] Ore detection module: used to set the optical path parameters of multiple optical paths according to the mining area data of the ore to be sorted area; adjust the parameters of the multiple optical paths based on real-time environmental data to obtain adjusted parameters; divide the mining area into several detection areas according to the mining area data; detect the several detection areas according to the adjusted parameters to obtain detection data;

[0009] Terminal sorting module: used to sort ore according to detection data.

[0010] Preferably, the setting of optical path parameters of multiple optical paths according to the mining area data of the ore to be sorted area includes:

[0011] Retrieve mining area data; the mining area data includes: the area to be sorted, the type of ore, and historical sorting data; integrate the ore type and historical sorting data in the mining area data into an optical path analysis sequence;

[0012] Call the parameter analysis model, input the optical path analysis sequence into the parameter analysis model, and obtain the parameter label; match the corresponding optical path parameters based on the parameter label; among which, the optical path parameters include: the number of optical paths, the wavelength of the optical path, the wavelength ratio, and the intensity of the optical path. The parameter analysis model is constructed based on the artificial intelligence model.

[0013] It should be noted that the historical sorting data includes: ore type, optical path parameters and misjudgment rate; parameter labels are set as positive integers; and the number of parameter labels is set according to the number of optical paths analyzed.

[0014] The present invention integrates the ore types and historical sorting data in the mining area data into an optical path analysis sequence; uses a parameter analysis model to analyze the optical path analysis sequence to obtain corresponding parameter labels, and matches the corresponding optical path parameters according to the parameter labels; using the model for analysis can quickly obtain the optical path parameters, which is conducive to improving the efficiency of data processing.

[0015] Preferably, the parameter analysis model is constructed based on an artificial intelligence model, including:

[0016] Select a suitable model and deep learning framework from several artificial intelligence models; define the model structure and set training parameters based on the selected deep learning framework to obtain a training model;

[0017] Acquire standard training data; wherein the standard training data includes: standard input data consistent with the content attributes of the optical path analysis sequence, and standard output data consistent with the content attributes of the parameter label;

[0018] The standard training data is divided into a training set, a validation set, and a test set according to a set ratio; the training model is trained using the training set; the internal parameters of the training model are adjusted using the validation set; the training model is tested using the test set to obtain the test index; when the test index is greater than the index threshold, the training model is marked as a parameter analysis model; otherwise, the training model is rebuilt and trained.

[0019] It should be noted that the test indicators include: accuracy, recall rate, F1 score and stability; the division ratio of training data is set by expert assessment; the indicator threshold is set based on actual experience. When the required optical path parameter accuracy is high, the indicator threshold is set higher; when the required optical path parameter accuracy is not high, the indicator threshold is set lower.

[0020] Preferably, the adjusting of the parameters of the multiple light paths based on the real-time environmental data includes:

[0021] Retrieve real-time environmental data, including real-time light intensity, real-time color temperature, and dust concentration;

[0022] Analyze the impact coefficient of real-time environmental data on the test results; match the impact coefficient with the impact step library to obtain the corresponding adjustment step; adjust the optical path parameters based on the adjustment step; test the adjusted optical path parameters to obtain the test results; when the test result is passed, mark the adjusted optical path parameters as adjustment parameters; otherwise, continue to adjust the optical path parameters.

[0023] The present invention adjusts the step size according to the influence coefficient of real-time environmental data on the detection result, and adjusts the optical path parameters according to the adjusted step size; and tests the adjusted optical path parameters. When the test passes, the adjusted optical path parameters are marked as adjustment parameters. The optical path data can be adjusted according to the influence of the real-time environmental data on the result of ore sorting, which is beneficial to reducing the interference of the real-time environment on the detection result and lowering the misjudgment rate of recognition.

[0024] Preferably, the analysis of the influence coefficient of real-time environmental data on the detection results includes:

[0025] Obtain historical detection data; where historical detection data includes: historical environmental data and historical misjudgment rate; the data type of historical environmental data is consistent with that of real-time environmental data;

[0026] A multivariate regression model is constructed: WP = β0 + β1GQ + β2SW + β3FN. The multivariate regression model is trained using historical detection data to obtain the influence coefficients. Among them, β1 is the influence coefficient of real-time light intensity; β2 is the influence coefficient of real-time color temperature; β3 is the influence coefficient of dust concentration; GQ, SW and FN represent real-time light intensity, real-time color temperature and dust concentration, respectively; β0 represents the error baseline value; and WP represents the misjudgment rate.

[0027] It should be noted that when using historical test data to train the multivariate regression model, the historical test data needs to be divided into a training set and a test set according to a certain ratio; the training set is used to train the multivariate regression model, and the test set is used to test the multivariate regression model. When the test meets the standard, the training is completed.

[0028] The present invention constructs a multiple regression model and uses historical detection data to train the multiple regression model to obtain corresponding influence coefficients; using historical detection data to train the multiple regression model can make the influence coefficients obtained by training more in line with reality, which is conducive to improving the accuracy of the detection results.

[0029] Preferably, the testing of the adjusted optical path parameters to obtain test results includes:

[0030] Retrieving the adjusted optical path parameters; irradiating the sorted ore with the adjusted optical path parameters, and collecting test data using a data sensor; wherein the test data includes: reflected light, transmitted light, and thermal radiation data;

[0031] Obtain the standard data of the corresponding ore; calculate the difference between the test data and the standard data, and when the absolute value of the difference is less than the difference threshold, mark the test result as passed; otherwise, mark the test result as failed.

[0032] It should be noted that the difference threshold is set according to actual requirements. When accurate identification of ore is required, the difference threshold is set to a smaller value; otherwise, the difference threshold is set to a larger value. In addition, during the test, the sorted ore is placed in the corresponding mining environment for testing; this can improve the accuracy of the test results.

[0033] The present invention utilizes adjusted optical path parameters to illuminate the sorted ore, utilizes a data sensor to collect test data, and analyzes whether the test data meets the standards based on standard data; it is capable of performing a preliminary test on the adjusted optical path parameters; it avoids a high misjudgment rate in ore sorting due to inadequate adjustment of the optical path parameters; and it is beneficial to improving the accuracy of ore identification.

[0034] Preferably, the mining area is divided into several detection areas according to the mining area data, including:

[0035] Retrieve the area to be sorted from the mining area data; divide the area to be sorted into several equal areas according to the set side length; identify the ores in the several equal areas based on the ore color to obtain several ore centers;

[0036] Mark the equally divided areas that do not contain the ore center to obtain the marked areas; calculate the distances from the marked areas to several ore centers, and select the equally divided areas corresponding to the ore center with the shortest distance as the merged areas; merge the marked areas with the merged areas to obtain several detection areas.

[0037] The present invention divides the area to be sorted into several equal areas, and identifies the several equal areas according to the color of the ore to obtain several ore centers, and divides the detection area according to the distance between the marked area and the ore center; each detection area contains ore, and the unidentified equal areas are merged with the area containing the ore center, which provides a basis for the subsequent accurate identification of the ore, avoids the occurrence of identification omissions, and is conducive to the analysis of ore yield.

[0038] Preferably, the detecting of a plurality of detection areas according to the adjustment parameters to obtain detection data includes:

[0039] Recall the adjustment parameters, divide the detection areas into a number of detection grids and number them JGij, and use the adjustment parameters to illuminate the detection areas; where JGij represents the detection grid in the i-th row and j-th column; i = 1, 2, ..., n, where n is a positive integer; j = 1, 2, ..., m, where m is a positive integer;

[0040] A data sensor is used to collect the irradiation data of several detection grids, wherein the irradiation data includes reflected light, transmitted light and thermal radiation data; a standard irradiation range of the ore is obtained; the irradiation data of the several detection grids are compared with the standard irradiation range, and the detection grids within the standard irradiation range are marked; the marked detection grids are integrated to obtain the detection data.

[0041] The present invention divides the detection area into several detection grids, illuminates the several detection areas with light by adjusting parameter settings, and uses data sensors to collect illumination data of the several detection grids; marks the several detection grids based on the standard illumination range and illumination data, and integrates the marked detection grids to obtain detection data; provides a data basis for subsequent sorting, and analyzes the data of the several detection grids to improve the accuracy of detection.

[0042] Preferably, the ore is sorted according to the detection data, including:

[0043] Retrieve the test data; construct a plan view of the mining area with sorting area; mark several test grids in the test data on the plan view to obtain the test mark area; integrate adjacent test mark areas to obtain the ore sorting area;

[0044] A sorting control signal is generated according to the ore sorting area, and the sorting control signal is sent to the corresponding ore sorting equipment to sort the ore.

[0045] A second aspect of the present invention provides an ore sorting method based on multi-optical path compensation, comprising:

[0046] Step S1: Acquire real-time environmental data of the ore separation area and mining area data;

[0047] Step S2: setting optical path parameters of the multi-optical path according to the mining area data of the ore to be sorted area; adjusting the parameters of the multi-optical path based on the real-time environmental data to obtain adjusted parameters;

[0048] Step S3: Divide the mining area into several detection areas according to the mining area data; detect the several detection areas according to the adjustment parameters to obtain detection data;

[0049] Step S4: sorting the ore according to the detection data.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] 1. The present invention integrates the ore types and historical sorting data in the mining area data into an optical path analysis sequence; uses a parameter analysis model to analyze the optical path analysis sequence to obtain corresponding parameter labels, and matches the corresponding optical path parameters according to the parameter labels; the optical path parameters can be quickly obtained by using the model analysis, which is beneficial to improving the efficiency of data processing; the step size is adjusted according to the influence coefficient of the real-time environmental data on the detection results, and the optical path parameters are adjusted according to the adjustment step size; and the adjusted optical path parameters are tested. When the test passes, the adjusted optical path parameters are marked as adjustment parameters; the optical path data can be adjusted according to the influence of the real-time environmental data on the results of ore sorting; which is beneficial to reducing The real-time environment interferes with the detection results, reducing the misjudgment rate of identification; constructing a multivariate regression model, and using historical detection data to train the multivariate regression model to obtain the corresponding influence coefficient; using historical detection data to train the multivariate regression model can make the training influence coefficient more in line with reality, which is beneficial to improving the accuracy of the detection results; using the adjusted optical path parameters to illuminate the sorted ore, and using data sensors to collect test data, and analyzing whether the test data meets the standards based on standard data; being able to conduct preliminary tests on the adjusted optical path parameters; avoiding a high misjudgment rate in ore sorting due to inadequate adjustment of the optical path parameters; and being beneficial to improving the accuracy of ore identification.

[0052] 2. The present invention divides the area to be sorted into several equal areas, and identifies the several equal areas according to the color of the ore to obtain several ore centers, and divides the detection area according to the distance between the marked area and the ore center; each detection area contains ore, and the unidentified equal areas are merged with the area containing the ore center, which provides a basis for the subsequent accurate identification of the ore, avoids the occurrence of identification omissions, and is conducive to the analysis of ore yield; the detection area is divided into several detection grids, and the several detection areas are illuminated by light using the adjustment parameter settings, and the illumination data of the several detection grids are collected using data sensors; the several detection grids are marked based on the standard illumination range and illumination data, and the marked detection grids are integrated to obtain detection data; a data basis is provided for subsequent sorting, and the data of the several detection grids are analyzed to improve the accuracy of detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. 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 any creative work.

[0054] Figure 1 A schematic diagram of the overall steps of the system of the present invention;

[0055] Figure 2 Schematic diagram of the optical path parameter setting and adjustment steps of the present invention;

[0056] Figure 3 Schematic diagram of the ore identification and sorting steps of the present invention;

[0057] Figure 4 Schematic diagram of the overall steps of the method of the present invention. DETAILED DESCRIPTION

[0058] The technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0059] See also Figure 1 , the first embodiment of the present invention provides an ore sorting system based on multi-optical path compensation, comprising: an ore detection module, and a data acquisition module connected thereto, and a terminal sorting module;

[0060] Data acquisition module: used to obtain real-time environmental data of the ore sorting area and mining area data;

[0061] Ore detection module: used to set the optical path parameters of multiple optical paths according to the mining area data of the ore to be sorted area; adjust the parameters of the multiple optical paths based on real-time environmental data to obtain adjusted parameters; divide the mining area into several detection areas according to the mining area data; detect the several detection areas according to the adjusted parameters to obtain detection data;

[0062] Terminal sorting module: used to sort ore according to detection data.

[0063] See also Figure 2 , obtain the mining area data of the area where the ore is to be sorted; wherein the mining area data includes: the area to be sorted, the type of ore and the historical sorting data; integrate the ore type and the historical sorting data in the mining area data into an optical path analysis sequence; call the parameter analysis model, input the optical path analysis sequence into the parameter analysis model, and obtain the parameter label; match the corresponding optical path parameters based on the parameter label; wherein the optical path parameters include: the number of optical paths, the wavelength of the optical path, the wavelength ratio and the intensity of the optical path, and the parameter analysis model is constructed based on the artificial intelligence model.

[0064] It should be noted that the historical sorting data includes: ore type, optical path parameters and misjudgment rate; parameter labels are set as positive integers; and the number of parameter labels is set according to the number of optical paths analyzed.

[0065] For example: suppose that hematite (Fe2O3) and quartz (SiO2) paragenetic ore is detected, and the area to be selected and the historical sorting data are integrated into an optical path analysis sequence; the parameter analysis model is called, and the optical path analysis sequence is input into the parameter analysis model to obtain the parameter labels: [3, (850, 650, 1550), (0.5, 0.3, 0.2), (100, 80, 120)]; then the corresponding optical path parameters matched according to the parameter labels are: 50% of the time, 850nm infrared with an intensity of 100 is used; 30% of the time, 650nm red visible light with an intensity of 80 is used; 20% of the time, 1550nm long-wave infrared with an intensity of 120 is used.

[0066] Acquire real-time environmental data of the ore sorting area; wherein the real-time environmental data includes: real-time light intensity, real-time color temperature and dust concentration; acquire historical detection data; wherein the historical detection data includes: historical environmental data and historical misjudgment rate; the data type of the historical environmental data is consistent with that of the real-time environmental data; construct a multiple regression model: WP = β0 + β1GQ + β2SW + β3FN; use the historical detection data to train the multiple regression model to obtain the influence coefficient; wherein β1 is the influence coefficient of real-time light intensity; β2 is the influence coefficient of real-time color temperature, and β3 is the influence coefficient of dust concentration; GQ, SW and FN represent real-time light intensity, real-time color temperature and dust concentration, respectively; β0 represents the error reference value; WP represents the misjudgment rate.

[0067] It should be noted that when using historical test data to train the multivariate regression model, the historical test data needs to be divided into a training set and a test set according to a certain ratio; the training set is used to train the multivariate regression model, and the test set is used to test the multivariate regression model. When the test meets the standard, the training is completed.

[0068] Match the influence coefficient with the influence step library to obtain the corresponding adjustment step; adjust the optical path parameters based on the adjustment step; use the adjusted optical path parameters to illuminate the sorted ore, and use the data sensor to collect test data; the test data includes: reflected light, transmitted light and thermal radiation data; obtain the standard data of the corresponding ore; calculate the difference between the test data and the standard data, and when the absolute value of the difference is less than the difference threshold, mark the test result as passed and mark the adjusted optical path parameters as adjusted parameters; otherwise, mark the test result as failed and continue to adjust the optical path parameters.

[0069] For example, assuming that the influence coefficient of real-time environmental data is obtained, the adjustment step size is obtained based on the influence coefficient analysis, and the optical path parameters are adjusted according to the adjustment step size;

[0070]

[0071] Assume that the real-time environmental data collected is: dust concentration is 10, real-time light intensity is 300, and real-time color temperature is 5000; then according to the adjustment measures, the irradiation time of 1550nm long-wave infrared is increased to 30%, and the intensity is adjusted to 115; the irradiation time of 850nm infrared is adjusted to 45%, and the intensity is adjusted to 95; the irradiation time of 650nm red visible light is adjusted to 25%, and the intensity is adjusted to 75.

[0072] It should be noted that the difference threshold is set according to actual requirements. When accurate identification of ore is required, the difference threshold is set to a smaller value; otherwise, the difference threshold is set to a larger value. In addition, during the test, the sorted ore is placed in the corresponding mining environment for testing; this can improve the accuracy of the test results.

[0073] It is worth noting that the parameter analysis model is built based on the artificial intelligence model, including:

[0074] Select a suitable model and deep learning framework from several artificial intelligence models; define the model structure and set training parameters based on the selected deep learning framework to obtain a training model;

[0075] Acquire standard training data; wherein the standard training data includes: standard input data consistent with the content attributes of the optical path analysis sequence, and standard output data consistent with the content attributes of the parameter label;

[0076] The standard training data is divided into a training set, a validation set, and a test set according to a set ratio; the training model is trained using the training set; the internal parameters of the training model are adjusted using the validation set; the training model is tested using the test set to obtain the test index; when the test index is greater than the index threshold, the training model is marked as a parameter analysis model; otherwise, the training model is rebuilt and trained.

[0077] It should be noted that the test indicators include: accuracy, recall rate, F1 score and stability; the division ratio of training data is set by expert assessment; the indicator threshold is set based on actual experience. When the required optical path parameter accuracy is high, the indicator threshold is set higher; when the required optical path parameter accuracy is not high, the indicator threshold is set lower.

[0078] See also Figure 3 , retrieve the area to be sorted in the mining area data; divide the area to be sorted into several equal areas according to the set side length; identify the ores in the several equal areas based on the ore color to obtain several ore centers; mark the equal areas that do not contain the ore center to obtain marked areas; calculate the distance from the marked area to the several ore centers, and select the equal areas corresponding to the ore centers with the shortest distance as the merged areas; merge the marked areas with the merged areas to obtain several detection areas.

[0079] Adjustment parameters are retrieved, and several detection areas are divided into several detection grids and numbered JGij. The adjustment parameters are used to illuminate the several detection areas; wherein JGij represents the detection grid in the i-th row and j-th column; i = 1, 2, ..., n, where n is a positive integer; j = 1, 2, ..., m, where m is a positive integer; illumination data of the several detection grids is collected using a data sensor; wherein the illumination data includes reflected light, transmitted light, and thermal radiation data; a standard illumination range of the ore is obtained; the illumination data of the several detection grids are compared with the standard illumination range, and the detection grids within the standard illumination range are marked; the marked detection grids are integrated to obtain the detection data.

[0080] Retrieve detection data; construct a plan view of the mining area with sorting area; mark several detection grids in the detection data on the plan view to obtain detection mark areas; integrate adjacent detection mark areas to obtain ore sorting areas; generate sorting control signals based on the ore sorting areas, and send the sorting control signals to corresponding ore sorting equipment to sort the ore.

[0081] It should be noted that when the mining area contains multiple types of ores, different types of ores are marked with different colors when marking the detection grid, and the detection mark areas of the same color are integrated when integrating; for example: assuming that JG23 in detection area 1 is detected to be hematite, the corresponding JG23 is marked in red, and JG47 in detection area 1 is quartz, the corresponding JG47 is marked in white.

[0082] See also Figure 4 The second embodiment of the present invention provides an ore sorting method based on multi-optical path compensation, comprising:

[0083] Step S1: Acquire real-time environmental data of the ore separation area and mining area data;

[0084] Step S2: setting optical path parameters of the multi-optical path according to the mining area data of the ore to be sorted area; adjusting the parameters of the multi-optical path based on the real-time environmental data to obtain adjusted parameters;

[0085] Step S3: Divide the mining area into several detection areas according to the mining area data; detect the several detection areas according to the adjustment parameters to obtain detection data;

[0086] Step S4: sorting the ore according to the detection data.

[0087] Some of the data in the above formula are calculated by removing the dimensions and taking their numerical values. The formula is a formula that is closest to the actual situation obtained by software simulation of a large amount of collected data; the preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained through simulation of a large amount of data.

[0088] The working principle of the present invention is as follows: the present invention obtains real-time environmental data and mining area data of the ore to be sorted area; sets optical path parameters of multiple optical paths according to the mining area data of the ore to be sorted area; adjusts the parameters of the multiple optical paths based on the real-time environmental data to obtain adjustment parameters; divides the mining area into several detection areas according to the mining area data; detects the several detection areas according to the adjustment parameters to obtain detection data; and sorts the ore according to the detection data.

[0089] The above embodiments are only used to illustrate the technical method of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. The ore sorting system based on multi-optical path compensation is characterized by: include: Ore detection module, and the connected data acquisition module and terminal sorting module; Data acquisition module: used to obtain real-time environmental data of the ore sorting area and mining area data; Ore detection module: used to set the optical path parameters of multiple optical paths according to the mining area data of the ore to be sorted; Adjust the parameters of the multiple optical paths based on real-time environmental data to obtain adjusted parameters; divide the mining area into several detection areas based on the mining area data; detect the several detection areas based on the adjusted parameters to obtain detection data; Terminal sorting module: used to sort ore according to detection data.

2. The ore sorting system based on multi-optical path compensation according to claim 1 is characterized in that: The optical path parameters of the multiple optical paths are set according to the mining area data of the ore to be sorted area, including: Retrieve mining area data; the mining area data includes: the area to be sorted, the type of ore, and historical sorting data; integrate the ore type and historical sorting data in the mining area data into an optical path analysis sequence; Call the parameter analysis model, input the optical path analysis sequence into the parameter analysis model, and obtain the parameter label; match the corresponding optical path parameters based on the parameter label; among which, the optical path parameters include: the number of optical paths, the wavelength of the optical path, the wavelength ratio, and the intensity of the optical path. The parameter analysis model is constructed based on the artificial intelligence model.

3. The ore sorting system based on multi-optical path compensation according to claim 2 is characterized in that: The parameter analysis model is constructed based on an artificial intelligence model and includes: Select a suitable model and deep learning framework from several artificial intelligence models; define the model structure and set training parameters based on the selected deep learning framework to obtain a training model; Acquire standard training data; wherein the standard training data includes: standard input data consistent with the content attributes of the optical path analysis sequence, and standard output data consistent with the content attributes of the parameter label; The standard training data is divided into a training set, a validation set, and a test set according to a set ratio; the training model is trained using the training set; the internal parameters of the training model are adjusted using the validation set; the training model is tested using the test set to obtain the test index; when the test index is greater than the index threshold, the training model is marked as a parameter analysis model; otherwise, the training model is rebuilt and trained.

4. The ore sorting system based on multi-optical path compensation according to claim 1 is characterized in that: The adjusting of the parameters of the multiple light paths based on the real-time environmental data includes: Retrieve real-time environmental data, including real-time light intensity, real-time color temperature, and dust concentration; Analyze the impact coefficient of real-time environmental data on the test results; match the impact coefficient with the impact step library to obtain the corresponding adjustment step; adjust the optical path parameters based on the adjustment step; test the adjusted optical path parameters to obtain the test results; when the test result is passed, mark the adjusted optical path parameters as adjustment parameters; otherwise, continue to adjust the optical path parameters.

5. The ore sorting system based on multi-optical path compensation according to claim 4 is characterized in that: The analysis of the influence coefficient of real-time environmental data on the detection results includes: Obtain historical detection data; where historical detection data includes: historical environmental data and historical misjudgment rate; the data type of historical environmental data is consistent with that of real-time environmental data; A multivariate regression model is constructed: WP = β0 + β1GQ + β2SW + β3FN. The multivariate regression model is trained using historical detection data to obtain the influence coefficients. Among them, β1 is the influence coefficient of real-time light intensity; β2 is the influence coefficient of real-time color temperature; β3 is the influence coefficient of dust concentration; GQ, SW and FN represent real-time light intensity, real-time color temperature and dust concentration, respectively; β0 represents the error baseline value; and WP represents the misjudgment rate.

6. The ore sorting system based on multi-optical path compensation according to claim 4 is characterized in that: The step of testing the adjusted optical path parameters to obtain test results includes: Retrieving the adjusted optical path parameters; irradiating the sorted ore with the adjusted optical path parameters, and collecting test data using a data sensor; wherein the test data includes: reflected light, transmitted light, and thermal radiation data; Obtain the standard data of the corresponding ore; calculate the difference between the test data and the standard data, and when the absolute value of the difference is less than the difference threshold, mark the test result as passed; otherwise, mark the test result as failed.

7. The ore sorting system based on multi-optical path compensation according to claim 1, characterized in that: The mining area is divided into several detection areas according to the mining area data, including: Retrieve the area to be sorted from the mining area data; divide the area to be sorted into several equal areas according to the set side length; identify the ores in the several equal areas based on the ore color to obtain several ore centers; Mark the equally divided areas that do not contain the ore center to obtain the marked areas; calculate the distances from the marked areas to several ore centers, and select the equally divided areas corresponding to the ore center with the shortest distance as the merged areas; merge the marked areas with the merged areas to obtain several detection areas.

8. The ore sorting system based on multi-optical path compensation according to claim 1, characterized in that: The detecting of a plurality of detection areas according to the adjustment parameters to obtain detection data includes: Recall the adjustment parameters, divide the detection areas into a number of detection grids and number them JGij, and use the adjustment parameters to illuminate the detection areas; where JGij represents the detection grid in the i-th row and j-th column; i = 1, 2, ..., n, where n is a positive integer; j = 1, 2, ..., m, where m is a positive integer; A data sensor is used to collect the irradiation data of several detection grids, wherein the irradiation data includes reflected light, transmitted light and thermal radiation data; a standard irradiation range of the ore is obtained; the irradiation data of the several detection grids are compared with the standard irradiation range, and the detection grids within the standard irradiation range are marked; the marked detection grids are integrated to obtain the detection data.

9. The ore sorting system based on multi-optical path compensation according to claim 1, characterized in that: The ore sorting according to the detection data includes: Retrieve the test data; construct a plan view of the mining area with sorting area; mark several test grids in the test data on the plan view to obtain the test mark area; integrate adjacent test mark areas to obtain the ore sorting area; A sorting control signal is generated according to the ore sorting area, and the sorting control signal is sent to the corresponding ore sorting equipment to sort the ore.

10. An ore sorting method based on multi-optical path compensation, applied to an ore sorting system based on multi-optical path compensation according to any one of claims 1 to 9, characterized in that: include: Step S1: Acquire real-time environmental data of the ore separation area and mining area data; Step S2: setting optical path parameters of the multi-optical path according to the mining area data of the ore to be sorted area; adjusting the parameters of the multi-optical path based on the real-time environmental data to obtain adjusted parameters; Step S3: Divide the mining area into several detection areas according to the mining area data; detect the several detection areas according to the adjustment parameters to obtain detection data; Step S4: sorting the ore according to the detection data.

Images