A method and system for locating personnel in a mine

Abstract

The present application relates to a kind of mine personnel positioning method and system, the method comprises: the current arbitrary one personnel in mine is determined as target personnel, obtains the multi-source positioning information of target personnel: baseline label positioning information, geomagnetic positioning information and pose positioning information;Obtain the multi-source positioning information corresponding to each time node in the preset historical time window of target personnel, generate adjacent multi-source feature vector;Adjacent multi-source feature vector is input into the decision contribution prediction model that is trained in advance, and the decision contribution factor sequence corresponding to multi-source positioning information is output, the decision contribution factor corresponding to each kind of positioning information in multi-source positioning information constitutes decision contribution factor sequenceBased on multi-source positioning information and its decision contribution factor sequence, input into the multi-source offset fusion algorithm model that is set in advance, output target personnel's target positioning information.Therefore, through the differentiation decision of different source positioning information, more accurate and reliable underground personnel positioning is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of mine safety monitoring, in particular to a mine personnel positioning method and system. BACKGROUND

[0002] Underground positioning is the basic guarantee for realizing underground safety production and supervision and management. At present, the underground positioning system commonly installed in mines mainly uses electromagnetic wave emission and signal recognition methods to realize underground positioning, such as RFID technology, Zigbee technology, PHS technology, WI-FI technology, etc. The positioning technology service object of this kind of method is mainly the management layer, which realizes the supervision and management of underground personnel or dynamic targets. Its positioning process depends on the underground power supply and communication system. Once the underground power supply or signal is unstable, the positioning device cannot work normally, and it cannot meet the requirements of emergency positioning and rescue positioning in emergency situations. In addition, the underground geomagnetic positioning method based on magnetic characteristics is difficult to adapt to complex mine environments due to the influence of iron and other metals, and the positioning accuracy is relatively large.

[0003] The Chinese invention patent with patent application number 202210024380.7 discloses a Baseline-RFMDR-based underground positioning method and underground positioning device. The first positioning of the underground personnel is obtained based on the included angle between the underground personnel and the mine roadway baseline tag. The second positioning of the underground personnel is obtained based on the geomagnetic data for geomagnetic matching. The third positioning of the underground personnel is obtained based on the walking average step length, the walking step number, and the walking direction. The first positioning, the second positioning, and the third positioning are fused and solved with the minimum error to obtain the accurate positioning of the underground personnel. Thus, the dependence on the underground power supply and communication system is small, and it is more suitable for complex mine environments.

[0004] However, in the actual underground positioning scene, if there are abnormal values (due to device failure, environmental interference, etc.) in the multiple positioning results, the fusion solving process may still be affected by these abnormal values, resulting in fluctuations, and easily falling into the global optimal solution of multiple positioning, which leads to the deviation of the accurate positioning result from the actual positioning without local tendency. SUMMARY

[0005] The present application provides a mine personnel positioning method and system to solve the technical problems in the prior art. Through the differentiated decision of different source positioning information, more accurate and reliable underground personnel positioning is realized.

[0006] The technical solution of the present application to solve the above technical problems is as follows:

[0007] A mine personnel positioning method, comprising:

[0008] S101, determine an arbitrary person in the current mine as a target person, obtain multi-source positioning information of the target person, the multi-source positioning information including three kinds of positioning information: baseline tag positioning information, geomagnetic positioning information and pose positioning information, and the types of the positioning information sources are respectively denoted as a, b and c;

[0009] S102, obtain multi-source positioning information corresponding to each time node of the target person in a preset historical time window, and generate a neighboring multi-source feature vector;

[0010] S103, input the neighboring multi-source feature vector into a decision contribution prediction model trained in advance, output a decision contribution factor sequence corresponding to the multi-source positioning information, and the decision contribution factors corresponding to each kind of positioning information in the multi-source positioning information constitute the decision contribution factor sequence;

[0011] S104, based on the multi-source positioning information and the decision contribution factor sequence thereof, input into a preset multi-source offset fusion algorithm model, and output target positioning information of the target person.

[0012] Preferably, the baseline tag positioning information is obtained by: obtaining a plurality of groups of baseline tag signals in the region, including tag coordinates and tag signal emission angles, selecting a group of baseline tags with the largest tag signal emission angle as a main baseline tag, based on the main baseline tag, selecting an adjacent group of baseline tags as a secondary baseline tag, and using the coordinate information of the main baseline tag and the secondary baseline tag and the tag signal emission angle to obtain the baseline tag positioning information of the target person.

[0013] Preferably, the neighboring multi-source feature vector is obtained by:

[0014] A1, in a preset historical time window, collect multi-source positioning information of each time node, including baseline tag positioning information, geomagnetic positioning information and pose positioning information;

[0015] A2, based on the baseline tag positioning information, the geomagnetic positioning information and the pose positioning information and the corresponding time nodes, generate a plurality of data points, each time node and the longitude in the positioning information constitute a data point, and perform fitting to obtain a trajectory curve in the preset historical time window, the horizontal axis of the trajectory curve represents the time node, and the vertical axis represents the longitude;

[0016] A3, based on the trajectory curve corresponding to each kind of positioning information in the multi-source positioning information, calculate the offset values between the trajectory curves of the other two kinds of positioning information respectively, and generate offset pairs;

[0017] A4, the offset pairs corresponding to the baseline tag positioning information, the geomagnetic positioning information and the pose positioning information respectively constitute the neighboring multi-source feature vector.

[0018] Preferably, the offset value is calculated according to the following formula:

[0019]

[0020] Wherein, D is the offset value between the trajectory curves of two kinds of positioning information, t1 is the starting time node of the historical time window, t2 is the terminal time node of the historical time window, is the longitude value of the trajectory curve of the positioning information of type a at time node t, is the longitude value of the trajectory curve of the positioning information of type b at time node t.

[0021] Preferably, the adjacent multi-source feature vector is denoted as [( , ), ( , ), ( , )] wherein ( , ) is the offset pair of baseline label positioning information, ( , ) is the offset pair of geomagnetic positioning information, ( , ) is the offset pair of pose positioning information, is the offset value between the trajectory curve of the a-type positioning information and the trajectory curve of the b-type positioning information, is the offset value between the trajectory curve of the a-type positioning information and the trajectory curve of the c-type positioning information.

[0022] Preferably, the obtaining manner of the pre-trained decision contribution prediction model comprises:

[0023] B1, collect a large number of adjacent multi-source feature vectors of personnel in a mine at historical time, and label each adjacent multi-source feature vector, and the labeling content is set as: decision contribution factor sequence; wherein, the labeling manner comprises: the sequence composed of the distance value proportion between the actual positioning information corresponding to the historical time and the baseline label positioning information, the geomagnetic positioning information and the pose positioning information corresponding thereto, as the decision contribution factor sequence of the adjacent multi-source feature vector;

[0024] B2, use the labeled historical adjacent multi-source feature vector as the training data set, train the pre-selected neural network structure using the training data set, optimize the model parameters, and obtain the final decision contribution prediction model.

[0025] Preferably, the pre-set multi-source offset fusion algorithm is:

[0026]

[0027] wherein, target positioning information of the target personnel, respectively, decision contribution factors of baseline tag positioning information, geomagnetic positioning information, and pose positioning information, is a distance value between the baseline tag positioning information and the target positioning information, is a distance value between the geomagnetic positioning information and the target positioning information, is a distance value between the pose positioning information and the target positioning information.

[0028] Preferably, after the S103, the method further comprises:

[0029] S201, obtaining underground mine image information collected by a tunnel monitoring device closest to a main baseline tag at a current time;

[0030] S202, inputting the underground mine image information, the shooting parameters of the monitoring device, and the coordinate information of the main baseline tag into a pre-trained mine target recognition model to obtain visual positioning information of the target personnel; wherein the mine target recognition model comprises a pixel classification model, a visual positioning prediction module, and a target personnel determination module, the underground mine image information is input into the pixel classification model to output a mine classification map, the mine classification map and the shooting parameters are input into the visual positioning prediction module to output all target objects and their visual positioning information, and the all target objects and their visual positioning information and the coordinate information of the main baseline tag are input into the target personnel determination module to output the visual positioning information of the target personnel;

[0031] S203, correcting the decision contribution factor sequence based on the relationship between the visual positioning information of the target personnel and the multi-source positioning information to obtain an updated decision contribution factor sequence to replace the original decision contribution factor sequence.

[0032] Preferably, in the S203, the decision contribution factor sequence is corrected, specifically including:

[0033] D1, a correction factor of each decision contribution factor is calculated according to the following formula:

[0034]

[0035]

[0036] wherein, i and j are used to represent the category number of the multi-source positioning information, including a, b, and c, is a distance value between the visual positioning information and the i-th multi-source positioning information, is a reliability degree value of the i-th multi-source positioning information, a contribution factor of a preliminary decision for the i-th multi-source positioning information, a distance value between the visual positioning information and the j-th multi-source positioning information, a reliability value of the j-th multi-source positioning information, a positive number parameter preset for controlling the speed of attenuation;

[0037] D2, generating a new sequence of decision contribution factors by taking the correction factor of each decision contribution factor as a new decision contribution factor.

[0038] A mine personnel positioning system, comprising: an acquisition module, a decision module and a fusion positioning module;

[0039] The acquisition module is configured to determine any one of the current mine personnel as a target personnel, obtain multi-source positioning information of the target personnel, and the multi-source positioning information comprises three kinds of positioning information: baseline tag positioning information, geomagnetic positioning information and pose positioning information, and the positioning information sources are respectively denoted as a, b and c; the acquisition module is configured to obtain the multi-source positioning information corresponding to each time node of the target personnel in a preset historical time window, and generate a neighboring multi-source feature vector;

[0040] The decision module is configured to input the neighboring multi-source feature vector into a pre-trained decision contribution prediction model, output a sequence of decision contribution factors corresponding to the multi-source positioning information, and each decision contribution factor corresponding to each kind of positioning information in the multi-source positioning information constitutes the sequence of decision contribution factors.

[0041] The fusion positioning module is configured to input the multi-source positioning information and the sequence of decision contribution factors into a pre-set multi-source offset fusion algorithm model, and output target positioning information of the target personnel.

[0042] The beneficial effects of the present application are:

[0043] The neighboring multi-source feature vector and the decision contribution prediction model are introduced to evaluate the contribution degree of each source positioning information, the multi-source offset fusion algorithm is adopted, the decision contribution factor is combined for calculation of the target positioning information, the accuracy and robustness of the positioning are improved, the influence of abnormal values on the positioning result is reduced by considering the relative consistency and difference of the multi-source positioning information, the abnormal values are processed by the decision contribution factor, the problem that the fusion calculation process may be affected by fluctuations when there are abnormal values in the multi-source fusion result is solved, and more accurate and reliable underground personnel positioning is achieved.

[0044] By introducing visual positioning information and combining multi-source positioning information for correction, the positioning result is more accurate; dynamically adjusting the decision contribution factor of positioning information of different sources, reducing the influence of abnormal value on the positioning result, making up for possible error defects of the model, and improving the robustness of positioning; through real-time acquisition of image information and calculation of correction factor, real-time positioning information update and correction can be supported. BRIEF DESCRIPTION OF DRAWINGS

[0045] Figure 1 A flowchart of a mine personnel positioning method according to an embodiment of the present application is shown in the figure.

[0046] Figure 2 A structure diagram of a mine personnel positioning system according to an embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0047] The technical solutions in the embodiments of the present application will be clearly and completely described in combination with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0048] In the description of the present application, the terms "first", "second" are used only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "multiple" is two or more, unless otherwise specifically limited.

[0049] In the description of the present application, the term "for example" is used to indicate "as an example, illustration or explanation". Any embodiment described as "for example" in the present application is not necessarily interpreted as more preferred or more advantageous than other embodiments. The following description is given in order to enable any person skilled in the art to implement and use the present application. In the following description, details are listed for the purpose of explanation. It should be understood that those skilled in the art can realize the present application without using these specific details. In other examples, well-known structures and processes will not be described in detail in order to avoid unnecessary details making the description of the present application obscure. Therefore, the present application is not intended to be limited to the shown embodiments, but is consistent with the broadest scope of principles and features disclosed in the present application.

[0050] Embodiment 1

[0051] Figure 1is a flowchart of a mine personnel positioning method according to an embodiment of the present application.

[0052] As shown in Figure 1 A mine personnel positioning method comprises the following steps:

[0053] S101, any one of the current mine personnel is determined as the target personnel, and the multi-source positioning information of the target personnel is obtained, the multi-source positioning information includes three kinds of positioning information: baseline tag positioning information, geomagnetic positioning information and pose positioning information, and the types of the positioning information sources are respectively denoted as a, b and c.

[0054] Specifically, each mine personnel carries a positioning device, and a plurality of groups of baseline tags with equal density are arranged equidistantly on the sidewall of the mine roadway, each group of baseline tags is symmetrically arranged on the two sides of the roadway, these tags are similar to the bar code or RFID tag in the supermarket, and are used for identifying a specific position in the roadway, the baseline tag includes tag coordinates, and is used for emitting a signal to be received by the positioning device carried by the personnel, and the baseline tag positioning information of the personnel is obtained in the following manner: the positioning device obtains the tag signal of a plurality of groups of baseline tags in the region, including the tag coordinates and the tag signal emission angle (the tag signal emission angle represents the included angle between the tag signal emission direction and the direction of the roadway, and is used for judging the relative position of the underground personnel and the tag), and a group of baseline tags with the largest tag signal emission angle is selected as the main baseline tag, which usually means that the relative position of the underground personnel and the tag is closest or the signal is strongest, based on the main baseline tag, a secondary baseline tag is further selected for assisting in calculating the position of the underground personnel, and the position coordinates of the underground personnel are calculated through geometric relationship by using the coordinate information of the main baseline tag and the secondary baseline tag and the tag signal emission angle.

[0055] The acquisition method of the geomagnetic positioning information is that the geomagnetic data of the underground personnel walking through the path in the mine is obtained, the geomagnetic data is specific data of the earth's magnetic field in the mine, which is similar to the GPS signal, but is not affected by the sky obstruction, the measured geomagnetic data is compared with the pre-stored geomagnetic map to determine the position of the underground personnel, and the geomagnetic map is pre-measured in the mine and contains the geomagnetic features of each place in the mine.

[0056] The acquisition method of the pose positioning information is that the walking average step length, the walking step number and the walking direction of the underground personnel in the mine are obtained based on the pose sensor arranged in the positioning device, the walking average step length is the average distance of each step when the underground personnel walks, the walking step number is the number of steps from the starting point to the current position, and the walking direction is the direction of the underground personnel walking (the included angle based on a certain direction), and the pose positioning information is obtained based on the walking data.

[0057] It should be noted that the detailed calculation and acquisition process of the baseline label positioning information, the geomagnetic positioning information and the pose positioning information can refer to related prior art, and the present application will not be repeated.

[0058] In S102, multi-source positioning information corresponding to each time node of the target person in a preset historical time window is acquired, and a neighboring multi-source feature vector is generated.

[0059] In some embodiments, in step S102, the neighboring multi-source feature vector is acquired in the following manner:

[0060] A1. In a preset historical time window (such as the past 2 minutes), multi-source positioning information of each time node (such as every second) is collected, including baseline label positioning information, geomagnetic positioning information and pose positioning information.

[0061] A2. Based on the baseline label positioning information, the geomagnetic positioning information and the pose positioning information and their corresponding time nodes, a plurality of data points are generated, each time node and the longitude (or latitude) in the positioning information constitute a data point, fitting (such as polynomial fitting) is performed, and a trajectory curve in the preset historical time window is obtained, the horizontal axis of the trajectory curve represents the time node, and the vertical axis represents the longitude (or latitude).

[0062] A3. Based on the trajectory curve of each kind of multi-source positioning information, the offset value between the trajectory curve of the other two kinds of positioning information is calculated respectively, and an offset pair is generated.

[0063] Specifically, the offset value is calculated according to the following formula:

[0064]

[0065] Wherein, D is the offset value between the trajectory curves of two kinds of positioning information, t1 is the starting time node of the historical time window, t2 is the terminal time node of the historical time window, is the trajectory curve function of the positioning information of type a, that is, the longitude value at time node t, is the trajectory curve function of the positioning information of type b, that is, the longitude value at time node t.

[0066] Specifically, the offset pair corresponding to a certain kind of positioning information includes the offset values between the trajectory curve of the positioning information of this kind and the trajectory curves of the other two kinds of positioning information, for example, the offset pair of the positioning information of type a is set as (Dab, Dba). , ), wherein, is the offset value between the trajectory curve of the positioning information of type a and the trajectory curve of the positioning information of type b, The offset value between the trajectory curve of type A positioning information and the trajectory curve of type C positioning information.

[0067] By comparing the offset values ​​between different types of location information, we can understand their correlation. If the offset values ​​between the two types of location information are small, it means that their correlation is high; if the offset values ​​are large, it means that their correlation is low. The neighboring multi-source feature vector can also indirectly reflect the accuracy of the location information. If the offset values ​​of a certain type of location information are large compared to the other two types of location information, then the accuracy of this location information may be low.

[0068] It should be noted that the calculation of offset pairs takes into account the overall offset of a certain type of positioning information relative to other types of positioning information within the time window. The neighboring multi-source feature vector reflects the relative consistency and differences between different types of positioning information that are closest to the current time.

[0069] A4. Combine the offset pairs corresponding to the baseline label positioning information, geomagnetic positioning information, and pose positioning information to form a nearest-neighbor multi-source feature vector, denoted as [( , ), ( , ), ( , )].

[0070] in,( , ) represents the offset pairs for baseline label positioning information. , ) represents the offset pair of geomagnetic positioning information, ( , ) represents the offset pair of pose positioning information.

[0071] S103, input the neighboring multi-source feature vectors into the pre-trained decision contribution prediction model, and output the decision contribution factor sequence corresponding to the multi-source positioning information. The decision contribution factor corresponding to each type of positioning information in the multi-source positioning information constitutes the decision contribution factor sequence.

[0072] In some embodiments, the pre-trained decision contribution prediction model is obtained in the following ways:

[0073] B1. Collect a large number of neighboring multi-source feature vectors of underground personnel at historical moments (select historical moments when the actual positioning information is judged as the standard (i.e., the positioning error is small and will not have an impact), and specifically select training samples based on expert experience), and label each neighboring multi-source feature vector with the label content set as: decision contribution factor sequence.

[0074] The specific methods of labeling include:

[0075] The distance value proportion (for example, the proportion of the distance between the actual positioning information and the baseline label positioning information in the total distance) between the actual positioning information corresponding to the historical moment and the corresponding baseline label positioning information, geomagnetic positioning information, pose positioning information respectively constitutes a sequence, which is the decision contribution factor sequence of the adjacent multi-source feature vector.

[0076] B2, all the labeled historical adjacent multi-source feature vectors are taken as a training data set, and a pre-selected neural network structure is trained using the training data set to optimize the model parameters and obtain a final decision contribution prediction model.

[0077] With the arrival of new observation data, the adjacent multi-source feature vectors within the historical time window can be continuously updated, which enables the model to dynamically adjust its prediction and decision strategy to adapt to changing environments and needs.

[0078] S104, based on the multi-source positioning information and its decision contribution factor sequence[ , , ], input into the pre-set multi-source offset fusion algorithm model, and output the target positioning information of the target person.

[0079] In some embodiments, the pre-set multi-source offset fusion algorithm is:

[0080]

[0081] wherein, is the target positioning information of the target person, is the decision contribution factor of the baseline label positioning information, geomagnetic positioning information, and pose positioning information respectively, is the distance value between the baseline label positioning information and the target positioning information, is the distance value between the geomagnetic positioning information and the target positioning information, is the distance value between the pose positioning information and the target positioning information. Thus, the multi-source offset fusion algorithm obtains the target positioning information by minimizing the weighted distance sum, and the decision contribution factor is applied as the weight value, reflecting the reliability and accuracy of various types of positioning information at the current moment.

[0082] In summary, the adjacent multi-source feature vector in the historical time window contains rich corresponding offsets in time sequence, and these features have a positive guiding effect on the final prediction result. The adjacent multi-source feature vector quantifies the consistency and difference of the positioning information in the historical time window by calculating the offset values between different types of positioning information (baseline label positioning, geomagnetic positioning, and pose positioning). This consistency and difference are important indicators for evaluating the reliability of various types of positioning information, which helps the model make more accurate judgments when fusing multiple types of positioning information. In the multi-source offset fusion algorithm model, the adjacent multi-source feature vector serves as the input feature of the decision contribution model, providing a basis for fusing different positioning information and adjusting the weights of different positioning information, thereby obtaining more accurate fused positioning results. By calculating the adjacent multi-source feature vector in real time and inputting it into the model, real-time updating and correction of positioning information can be supported, which helps maintain the accuracy and reliability of positioning information in dynamic environments and reduces the impact of abnormal values or sudden disturbances on positioning results, improving the robustness of positioning.

[0083] The adjacent multi-source feature vector and the decision contribution prediction model are introduced to evaluate the contribution degree of each source of positioning information. The multi-source offset fusion algorithm is adopted to calculate the target positioning information combined with the decision contribution factor, which improves the accuracy and robustness of positioning. By considering the relative consistency and difference of multi-source positioning information, the impact of abnormal values on positioning results is reduced. The decision contribution factor is used to reduce the weight of current possible abnormal values, solving the problem that the fusion calculation process may be affected by fluctuations and abnormal values in the multi-positioning fusion result, achieving more accurate and reliable underground personnel positioning.

[0084] Embodiment 2

[0085] In some embodiments, after step S103, the method further comprises:

[0086] S201, obtaining mine underground image information collected by the nearest roadway monitoring device to the main baseline label at the current time.

[0087] In the mine roadway, monitoring devices (such as cameras) are arranged at key positions to monitor the activities underground. The nearest monitoring device to the main baseline label is selected to obtain the mine underground image information at the current time. The main baseline label is determined in step S101 and is used to identify a specific position in the roadway. By obtaining the image of the nearest monitoring device, it can be ensured that the image information is closely related to the position of the target personnel.

[0088] S202, inputting the mine underground image information, the shooting parameters of the monitoring device, and the coordinate information of the main baseline label into the pre-trained mine target recognition model to obtain the visual positioning information of the target personnel.

[0089] Specifically, the mine target recognition model comprises a pixel classification model, a visual positioning prediction module and a target personnel determination module. The mine image information is input into the pixel classification model, and a mine classification map is output. The mine classification map and the shooting parameters are input into the visual positioning prediction module, and all target objects and their visual positioning information are output. The target objects and their visual positioning information, the main baseline label coordinate information are input into the target personnel determination module, and the visual positioning information of the target personnel is output.

[0090] The target objects can include mine cars, personnel and other landmark objects in the mine, and the shooting parameters include the focal length, optical center, tilt angle, horizontal field of view angle, vertical field of view angle, camera height and the like of the monitoring device camera.

[0091] The pixel classification module is configured to output a mine classification map with the same resolution as the input mine image information, and the classification value of each pixel represents the target object identifier to which the pixel in the input mine image information belongs. For the training of the pixel classification model, a large amount of historical mine image information is labeled to obtain a training data set, and the pre-set neural network structure is trained using the training data set to obtain the final pixel classification model.

[0092] Specifically, the visual positioning prediction module is specifically configured to:

[0093] C1, based on all pixel points in the mine classification map and their labeled target object identifiers, the pixel points are merged according to the target object identifiers (the pixel points with the same target object identifier and the discrete degree less than the discrete threshold are merged), at least one target object class is obtained, each target object class corresponds to a target object identifier, the original mine classification map is updated, and the latest mine classification map includes at least one target object class.

[0094] C2, the pixel points in the target object class are converted from the image coordinate system to the visual positioning information in the same coordinate system (i.e. scene coordinate system) as the position information corresponding to the main baseline label, specifically including: using camera calibration technology to obtain the intrinsic parameters (such as focal length, optical center, etc.) and extrinsic parameters (such as rotation matrix, translation vector) of the camera, and applying geometric transformation formula to convert image coordinates to actual scene coordinates:

[0095] The positioning information of each sub-pixel point is converted from the image coordinate system to the scene coordinate system through image scaling and shooting parameters. The coordinate information of the pixel points of the target object class in the image coordinate system is determined according to the camera parameters (focal length, light center position and distortion coefficient), and the image scaling is set in advance according to the camera parameters and the scene coordinate system, which is used to convert the pixel unit in the image into the physical unit in the real world. The determination of the image scaling usually depends on the camera calibration process, which usually involves the use of camera parameters, and the present application will not be repeated here. For each pixel point in the target object class, the image coordinates and the camera's internal and external parameters are used for conversion. The conversion formula usually involves converting the image coordinates into three-dimensional coordinates in the camera coordinate system, and then converting to the scene coordinate system through the external parameters. The specific conversion formula may differ depending on the camera model and the calibration method, and the present application will not be repeated here. Please refer to the existing related technology.

[0096] It should be noted that the main baseline tag coordinate information provides a known reference point in the scene, and the main baseline tag coordinate information and the monitoring device coordinate information (i.e. the position information of the camera) can be used to further correct and verify the conversion result. By comparing the relative position relationship between the converted coordinates and the main baseline tag coordinates, the parameters in the conversion formula can be adjusted to improve the accuracy of the conversion.

[0097] Specifically, the target personnel determination module is specifically configured to:

[0098] Filtering out the visual positioning information of the target object whose identification is personnel, and comparing the visual positioning information with the multi-source positioning information and the main baseline tag coordinate information, and taking the visual positioning information that meets the preset condition as the visual positioning information of the target personnel.

[0099] The preset condition is set as: under the premise that the distance value between the visual positioning information and all the baseline tag coordinate information received by the target personnel is the smallest, the distance value between the visual positioning information and the multi-source positioning information is the smallest (the distance value between the visual positioning information and each kind of positioning information can be calculated respectively, and then weighted and averaged based on the corresponding decision contribution factor). It should be noted that all the baseline tag coordinate information appears in pairs, and the distance value between the visual positioning information and all the baseline tag coordinate information received by the target personnel is set as the sum of the distance values between each coordinate information in each group of baseline tag coordinate information and the visual positioning information.

[0100] S203, based on the relationship between the visual positioning information of the target personnel and the multi-source positioning information, the decision contribution factor sequence is modified to obtain an updated decision contribution factor sequence to replace the original decision contribution factor sequence.

[0101] The modification of the decision contribution factor sequence specifically includes:

[0102] D1, the correction factor of each decision contribution factor is calculated according to the following formula:

[0103]

[0104]

[0105] Wherein, i and j are used to represent the category number of multi-source positioning information, including a, b, c, is the distance value between the visual positioning information and the i-th multi-source positioning information, is the reliability value of the i-th multi-source positioning information, is the original decision contribution factor of the i-th multi-source positioning information, is the distance value between the visual positioning information and the j-th multi-source positioning information, is the reliability value of the j-th multi-source positioning information, is a Gaussian function used to attenuate the consistency of positioning information according to distance, is a positive number parameter set in advance, used to control the speed of attenuation.

[0106] D2, the correction factor of each decision contribution factor is taken as a new decision contribution factor, and a new decision contribution factor sequence is generated.

[0107] In summary, by introducing visual positioning information and combining multi-source positioning information for correction, the visual positioning information provides intuitive image basis, which helps to reduce the error of other positioning information, so that the positioning result is more accurate; dynamically adjusting the decision contribution factor of positioning information of different sources and categories reduces the influence of abnormal values on the positioning result, makes up for possible error defects of the model, and improves the robustness of positioning; by real-time acquisition of image information and calculation of correction factor, real-time positioning information update and correction can be supported, which adapts to the complex and changeable mine environment.

[0108] Embodiment 3

[0109] Figure 2 is a structural schematic diagram of a mine personnel positioning system according to an embodiment of the present application.

[0110] As Figure 2 shown, a mine personnel positioning system comprises an acquisition module, a decision module and a fusion positioning module.

[0111] Specifically, the acquisition module is configured to determine any one person under the current mine as a target person, obtain multi-source positioning information of the target person, and the multi-source positioning information comprises three kinds of positioning information: baseline tag positioning information, geomagnetic positioning information and pose positioning information, and the positioning information sources are respectively denoted as a, b and c; the acquisition module is configured to obtain multi-source positioning information corresponding to each time node in a preset historical time window of the target person, and generate a neighboring multi-source feature vector;

[0112] The decision module is configured to input the neighboring multi-source feature vector into a decision contribution prediction model trained in advance, and output a decision contribution factor sequence corresponding to the multi-source positioning information, wherein a decision contribution factor corresponding to each kind of positioning information in the multi-source positioning information constitutes the decision contribution factor sequence.

[0113] The fusion positioning module is configured to input the multi-source positioning information and the decision contribution factor sequence thereof into a multi-source offset fusion algorithm model set in advance, and output target positioning information of the target person.

[0114] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0115] Those skilled in the art should understand that embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0116] The present application is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing devices to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices produce a device that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one flow or multiple flows and / or blocks Figure 1 The functions specified in one flow or multiple flows and / or blocks

[0117] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the Figure 1 function specified in the flow or flows and / or blocks Figure 1 of the block or blocks.

[0118] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions that are executed on the computer or other programmable apparatus provide steps for implementing the Figure 1 function specified in the flow or flows and / or blocks Figure 1 of the block or blocks.

[0119] While the preferred embodiments of the application have been described, additional variations and modifications can be made to the embodiments by those skilled in the art once they learn of the basic inventive concepts. Therefore, the appended claims are intended to cover all such additional variations and modifications as fall within the scope of the present application. What is claimed is:

[0120] Obviously, numerous modifications and variations of the present application are possible in light of the above teachings. It is therefore to be understood that within the scope of the appended claims and their equivalents, the application can be practiced otherwise than as specifically described.

Claims

1. A method of locating personnel underground in a mine, characterised by, The method comprises the steps of: S101, determining any one person in the current coal mine as a target person, obtaining multi-source positioning information of the target person, and the multi-source positioning information comprises three types of positioning information: baseline tag positioning information, geomagnetic positioning information and pose positioning information, and the types of the positioning information are respectively denoted as a, b and c; S102, obtaining multi-source positioning information corresponding to each time node of the target person in a preset historical time window, and generating a neighboring multi-source feature vector; S103, inputting the neighboring multi-source feature vector into a pre-trained decision contribution prediction model, and outputting a decision contribution factor sequence corresponding to the multi-source positioning information, wherein each type of positioning information in the multi-source positioning information corresponds to a decision contribution factor, and the decision contribution factors constitute the decision contribution factor sequence; S104, inputting the multi-source positioning information and the decision contribution factor sequence into a pre-set multi-source offset fusion algorithm model, and outputting target positioning information of the target person; The method comprises the steps of: A1, collecting multi-source positioning information of each time node in a preset historical time window, including baseline tag positioning information, geomagnetic positioning information and pose positioning information; A2, generating a plurality of data points based on the baseline tag positioning information, the geomagnetic positioning information and the pose positioning information and the corresponding time nodes, wherein the longitude of each time node and the positioning information constitutes a data point, fitting is performed to obtain a trajectory curve in the preset historical time window, the horizontal axis of the trajectory curve represents the time node, and the vertical axis represents the longitude; A3, calculating the offset values between each type of positioning information in the multi-source positioning information and the other two types of positioning information trajectory curves respectively based on the trajectory curves corresponding to each type of positioning information, and generating offset pairs; A4, combining the offset pairs corresponding to the baseline tag positioning information, the geomagnetic positioning information and the pose positioning information to form a neighboring multi-source feature vector.

2. The mine personnel positioning method according to claim 1, characterized in that, The baseline tag positioning information is obtained by the following method: obtaining a plurality of groups of baseline tag signals in the region, including tag coordinates and tag signal emission angles, selecting a group of baseline tags with the largest tag signal emission angle as a main baseline tag, selecting an adjacent group of baseline tags as a secondary baseline tag based on the main baseline tag, and obtaining the baseline tag positioning information of the target person by using the coordinate information of the main baseline tag and the secondary baseline tag and the tag signal emission angles.

3. The mine personnel positioning method of claim 1, wherein, The offset value is calculated according to the following formula: wherein D is an offset value between the trajectory curves of the two kinds of positioning information, t1 is a starting time node of a historical time window, t2 is a terminal time node of the historical time window, is a longitude value of the trajectory curve of the positioning information of kind a at the time node t, is a longitude value of the trajectory curve of the positioning information of kind b at the time node t.

4. The mine personnel positioning method of claim 1, wherein, The adjacent multi-source feature vector is denoted as [( , ), ( , ), ( , ) ], wherein (( , ) is an offset pair of baseline label positioning information, (( , ) is an offset pair of geomagnetic positioning information, (( , ) is an offset pair of pose positioning information, is an offset value between the trajectory curve of a kind of positioning information and the trajectory curve of b kind of positioning information, is an offset value between the trajectory curve of a kind of positioning information and the trajectory curve of c kind of positioning information.

5. The mine personnel positioning method of claim 1, wherein, The pre-trained decision contribution prediction model is obtained by the following method: B1, collecting a large number of neighboring multi-source feature vectors of personnel in the mine at historical time points, and labeling each neighboring multi-source feature vector, wherein the labeling content is set as a decision contribution factor sequence; wherein the labeling method comprises the following steps: a sequence composed of distance values between the actual positioning information corresponding to the historical time and the baseline tag positioning information, the geomagnetic positioning information and the pose positioning information, as the decision contribution factor sequence of the neighboring multi-source feature vector. B2, use the labeled historical adjacent multi-source feature vector as a training data set, train a pre-selected neural network structure using the training data set, optimize the model parameters, and obtain a final decision contribution prediction model.

6. The mine personnel positioning method of claim 1, wherein, The pre-set multi-source offset fusion algorithm is: wherein, a target positioning information of the target person, a decision contribution factor of the baseline tag positioning information, the geomagnetic positioning information, and the pose positioning information, respectively, a distance value between the baseline tag positioning information and the target positioning information, a distance value between the geomagnetic positioning information and the target positioning information, a distance value between the pose positioning information and the target positioning information.

7. The mine personnel positioning method of claim 2, wherein, After the S103, the method further comprises: S201, obtaining mine image information collected by a nearest roadway monitoring device to a main baseline label at a current time; S202, inputting the mine image information, the shooting parameters of the monitoring device, and the coordinate information of the main baseline label into a pre-trained mine target recognition model to obtain visual positioning information of the target personnel; wherein the mine target recognition model comprises a pixel classification model, a visual positioning prediction module, and a target personnel determination module, the mine image information is input into the pixel classification model, and a mine classification map is output; the mine classification map and the shooting parameters are input into the visual positioning prediction module, and all target objects and their visual positioning information are output; all target objects and their visual positioning information, and the coordinate information of the main baseline label are input into the target personnel determination module, and the visual positioning information of the target personnel is output; S203, based on the relationship between the visual positioning information of the target personnel and the multi-source positioning information, the decision contribution factor sequence is corrected to obtain an updated decision contribution factor sequence to replace the original decision contribution factor sequence.

8. A mine personnel location method according to claim 7, characterised in that, In the S203, the decision contribution factor sequence is corrected, specifically including: D1, the correction factor of each decision contribution factor is calculated according to the following formula: wherein i and j are used to represent the category number of multi-source positioning information, including a, b, c, is a distance value between the visual positioning information and the i-th multi-source positioning information, is a reliability degree value of the i-th multi-source positioning information, is an original decision contribution factor of the i-th multi-source positioning information, is a distance value between the visual positioning information and the j-th multi-source positioning information, is a reliability degree value of the j-th multi-source positioning information, is a positive number parameter preset for controlling the speed of attenuation; D2, the correction factor of each decision contribution factor is used as a new decision contribution factor to generate a new decision contribution factor sequence.

9. A mine personnel positioning system for use in the mine personnel positioning method according to any one of claims 1 to 8, characterized by Including: An acquisition module, a decision module, and a fusion positioning module; The acquisition module is used to determine any one personnel in the current mine as a target personnel, obtain multi-source positioning information of the target personnel, and the multi-source positioning information includes three kinds of positioning information: baseline label positioning information, geomagnetic positioning information, and pose positioning information, and the positioning information sources are respectively marked as a, b, and c; the acquisition module is used to obtain the multi-source positioning information corresponding to each time node of the target personnel in a preset historical time window, and generate an adjacent multi-source feature vector; The decision module is used to input the adjacent multi-source feature vector into a pre-trained decision contribution prediction model, output a decision contribution factor sequence corresponding to the multi-source positioning information, and each kind of positioning information in the multi-source positioning information corresponds to a decision contribution factor to form the decision contribution factor sequence; The fusion positioning module is used to input the multi-source positioning information and its decision contribution factor sequence into a pre-set multi-source offset fusion algorithm model to output target positioning information of the target personnel.

Images