Palmprint data recognition method and system based on deep learning

Abstract

The present invention provides a palm print data recognition method and system based on deep learning; the palm of the hand when five fingers are stretched out is detected by alternately integrating near-infrared absorption and visible light reflection, so as to obtain basic palm print information data and basic palm vein detection data; a palm print and palm vein basic information database is established according to the basic palm print information data and basic palm vein detection data, and palm print information data features and palm vein data features are extracted to construct a palm print and palm vein integrated recognition deep learning model; according to the palm print and palm vein basic information database, a palm print and palm vein integrated recognition deep learning model is trained to obtain a palm print and palm vein integrated recognition deep learning precision model; palm print information data and palm vein detection data are detected in real time and input into the palm print and palm vein integrated recognition deep learning precision model, palm print and palm vein integrated recognition results are output, and when any palm print and palm vein recognition item is abnormal, a palm print and palm vein recognition abnormality alarm is triggered.

Description

Technical Field

[0001] The present invention relates to the technical field of optical ray detection intelligent recognition, and more specifically, to a palmprint data recognition method and system based on deep learning. Background Art

[0002] Palmprint recognition and palm vein recognition are very important for the rapid development of intelligent recognition; palmprint recognition of palm texture uses the difference in the absorption rate of deoxyhemoglobin in the vein and other physiological tissues to near-infrared light to detect vein data; currently there are problems such as skin surface damage, aging and other factors affecting the accuracy of detection and recognition; specifically including: how to detect the palms in various states and accurately obtain basic palmprint information data and basic palm vein detection data, how to obtain accurate palmprint information data features and palm vein data features and build a palmprint and palm vein integrated recognition model, how to train the model and make palmprint and palm vein recognition more accurate, how to detect palmprint information data and palm vein detection data in real time and accurately output palmprint and palm vein integrated recognition results and promptly warn when palmprint and palm vein recognition anomalies are found, and other issues remain to be resolved; therefore, it is necessary to propose a palmprint data recognition method and system based on deep learning to at least partially solve the problems existing in the prior art. Summary of the invention

[0003] A series of simplified concepts are introduced in the content of the invention, which will be further described in detail in the specific implementation method section; the content of the invention of the present invention does not mean to attempt to limit the key features and essential technical features of the technical solution claimed for protection, nor does it mean to attempt to determine the scope of protection of the technical solution claimed for protection.

[0004] In order to at least partially solve the above problems, the present invention provides a palmprint data recognition method based on deep learning, comprising:

[0005] S100, detects the palm of the hand when five fingers are stretched out by alternately absorbing near-infrared light and reflecting visible light, and obtains basic palm print information data and basic palm vein detection data;

[0006] S200, establishing a palm print and palm vein basic information database based on basic palm print information data and basic palm vein detection data, extracting palm print information data features and palm vein data features, and constructing a palm print and palm vein integrated recognition deep learning model;

[0007] S300, training a palm print and palm vein integrated recognition deep learning model based on a palm print and palm vein basic information database, and obtaining a palm print and palm vein integrated recognition deep learning accurate model;

[0008] S400, real-time detection of palm print information data and palm vein detection data and input into the palm print and palm vein integrated recognition deep learning precision model, output the palm print and palm vein integrated recognition result, and trigger the palm print and palm vein recognition abnormality alarm when any palm print and palm vein recognition is abnormal.

[0009] Preferably, S100 includes:

[0010] S101, adopting near-infrared absorption and visible light reflection dual-light alternating integrated detection, setting a near-infrared absorption dual-light alternating integrated detector;

[0011] S102, alternately irradiating the palm with five fingers stretched out with visible light and near-infrared rays of a near-infrared absorption dual-light alternating integrated detector, and obtaining basic palm print information data and basic palm vein detection data through visible light ray reflection sensing and near-infrared ray reflection sensing detection.

[0012] Preferably, S200 includes:

[0013] S201, establishing a palm print and palm vein basic information database based on basic palm print information data and basic palm vein detection data;

[0014] S202, extracting palm print information data features and palm vein data features according to the palm print and palm vein basic information database;

[0015] S203, performing palm print and palm vein integrated identification through statistical analysis of palm print information data features and palm vein data features, and constructing a palm print and palm vein integrated identification deep learning model.

[0016] Preferably, S300 includes:

[0017] S301, establishing a palm print and palm vein training set, a palm print and palm vein verification set, and a palm print and palm vein test set based on a palm print and palm vein basic information database, and training a palm print and palm vein integrated recognition deep learning model;

[0018] S302, obtaining a precise palm print and palm vein integrated recognition deep learning model by training the palm print and palm vein integrated recognition deep learning model until the output of the palm print and palm vein integrated recognition deep learning model reaches a preset data accuracy.

[0019] Preferably, S400 includes:

[0020] S401, real-time detection of palm print information data and palm vein detection data, open the palm naturally, place the palm 5-10 cm above the sensor, and shake it slightly; obtain real-time detection of palm print information data and real-time detection of palm vein detection data;

[0021] S402, inputting the real-time palm print information data and the real-time palm vein detection data into the palm print and palm vein integrated recognition deep learning precision model;

[0022] S403, the palm print and palm vein integrated recognition deep learning precision model outputs the palm print and palm vein integrated recognition result;

[0023] S404, triggering a palm print and palm vein recognition abnormality alarm when any one of the palm print and palm vein recognition is abnormal according to the palm print and palm vein integrated recognition result.

[0024] The present invention provides a palmprint data recognition system based on deep learning, comprising:

[0025] The palm print and palm vein basic detection module detects the palm of the hand with five fingers stretched out by alternately absorbing near-infrared light and reflecting visible light, and obtains basic palm print information data and basic palm vein detection data;

[0026] The palm print and palm vein integrated recognition architecture module establishes a palm print and palm vein basic information database based on basic palm print information data and basic palm vein detection data, extracts palm print information data features and palm vein data features, and constructs a palm print and palm vein integrated recognition deep learning model;

[0027] Palm print and palm vein recognition model training module, based on the palm print and palm vein basic information database, trains a palm print and palm vein integrated recognition deep learning model to obtain a palm print and palm vein integrated recognition deep learning accurate model;

[0028] The real-time integrated palm print and palm vein recognition module detects palm print information data and palm vein detection data in real time and inputs them into the palm print and palm vein integrated recognition deep learning precision model, outputs the palm print and palm vein integrated recognition results, and triggers the palm print and palm vein recognition abnormality alarm when any of the palm print and palm vein recognition is abnormal.

[0029] Preferably, the palm print and palm vein basic detection module includes:

[0030] The dual-light alternating integrated setting unit adopts near-infrared absorption and visible light reflection dual-light alternating integrated detection, and sets a near-infrared absorption dual-light alternating integrated detector;

[0031] The absorption and reflection dual-light alternating detection unit uses visible light and near-infrared rays of the near-infrared absorption dual-light alternating integrated detector to alternately irradiate the palm when the five fingers are stretched out, and obtains basic palm print information data and basic palm vein detection data through visible light reflection sensing and near-infrared reflection sensing detection.

[0032] Preferably, the palm print and palm vein integrated recognition architecture module includes:

[0033] A palm print and palm vein basic information database unit is used to establish a palm print and palm vein basic information database based on basic palm print information data and basic palm vein detection data;

[0034] The information data feature extraction unit extracts the palm print information data features and palm vein data features according to the palm print and palm vein basic information database;

[0035] The integrated recognition deep learning model unit performs palm print and palm vein integrated judgment through statistical analysis of palm print information data features and palm vein data features, and constructs a palm print and palm vein integrated recognition deep learning model.

[0036] Preferably, the palm print and palm vein recognition model training module includes:

[0037] The deep learning model training unit establishes a palm print and palm vein training set, a palm print and palm vein verification set, and a palm print and palm vein test set based on the palm print and palm vein basic information database, and trains a palm print and palm vein integrated recognition deep learning model;

[0038] The deep learning precision model unit obtains the deep learning precision model of palm print and palm vein integrated recognition by training the deep learning model of palm print and palm vein integrated recognition until the output of the deep learning model of palm print and palm vein integrated recognition reaches the preset data accuracy.

[0039] Preferably, the palm print and palm vein real-time integrated recognition module includes:

[0040] Palm print and palm vein real-time detection unit, real-time detection of palm print information data and palm vein detection data, open your palm naturally, place your palm 5-10 cm above the sensor, and swing it slightly; obtain real-time detection of palm print information data and real-time detection of palm vein detection data;

[0041] A detection data input unit, which inputs real-time palm print information data and real-time palm vein detection data into a palm print and palm vein integrated recognition deep learning precision model;

[0042] Integrated recognition deep learning output unit: the palm print and palm vein integrated recognition deep learning precision model outputs the palm print and palm vein integrated recognition results;

[0043] The palm print and palm vein recognition abnormality alarm unit triggers a palm print and palm vein recognition abnormality alarm according to the palm print and palm vein integrated recognition result when any of the palm print and palm vein recognition is abnormal.

[0044] The beneficial effects of the above technical solution include:

[0045] The present invention provides a palm print data recognition method and system based on deep learning, which detects the palm of the five fingers when they are stretched out by alternately integrating near-infrared absorption and visible light reflection, and obtains basic palm print information data and basic palm vein detection data; a palm print and palm vein basic information database is established according to the basic palm print information data and basic palm vein detection data, and palm print information data features and palm vein data features are extracted to construct a palm print and palm vein integrated recognition deep learning model; according to the palm print and palm vein basic information database, a palm print and palm vein integrated recognition deep learning model is trained to obtain a palm print and palm vein integrated recognition deep learning precision model; palm print information data and palm vein detection data are detected in real time and input into the palm print and palm vein integrated recognition deep learning precision model, and a palm print and palm vein integrated recognition result is output, and when any palm print and palm vein recognition item is abnormal, a palm print and palm vein recognition abnormality is triggered. It can detect palm texture features in various states and accurately obtain basic palm print information data and basic palm vein detection data; it can obtain accurate palm print information data features and palm vein data features and construct a palm print and palm vein integrated recognition model, and the model accuracy is significantly improved; it can train the model and make palm print and palm vein recognition more accurate; it can detect palm print information data and palm vein detection data in real time and accurately output palm print and palm vein integrated recognition results, and it can promptly detect palm print and palm vein recognition anomalies and give timely alarms, and the real-time accuracy of palm print information and palm vein detection is significantly improved.

[0046] The present invention provides a palmprint data recognition method and system based on deep learning. Other advantages, objectives and features of the present invention will be partially reflected in the following description, and will also be understood by technicians in this field through research and practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0048] Figure 1 This is a diagram of an embodiment of a palmprint data recognition system based on deep learning of the present invention.

[0049] Figure 2 This is a diagram of an embodiment of a palmprint data recognition system product based on deep learning of the present invention.

[0050] Figure 3 This is a diagram of an embodiment of a palmprint data recognition method and system application scenario based on deep learning of the present invention. DETAILED DESCRIPTION

[0051] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments so that those skilled in the art can implement the invention according to the instructions. Figure 1-3 As shown, the present invention provides a palmprint data recognition method based on deep learning, comprising:

[0052] S100, detects the palm of the hand when five fingers are stretched out by alternately absorbing near-infrared light and reflecting visible light, and obtains basic palm print information data and basic palm vein detection data;

[0053] S200, establishing a palm print and palm vein basic information database based on basic palm print information data and basic palm vein detection data, extracting palm print information data features and palm vein data features, and constructing a palm print and palm vein integrated recognition deep learning model;

[0054] S300, training a palm print and palm vein integrated recognition deep learning model based on a palm print and palm vein basic information database, and obtaining a palm print and palm vein integrated recognition deep learning accurate model;

[0055] S400, real-time detection of palm print information data and palm vein detection data and input into the palm print and palm vein integrated recognition deep learning precision model, output the palm print and palm vein integrated recognition result, and trigger the palm print and palm vein recognition abnormality alarm when any palm print and palm vein recognition is abnormal.

[0056] The principle and effect of the above technical solution are as follows: the present invention provides a palmprint data recognition method based on deep learning, including: detecting the palm of the hand when five fingers are stretched out by alternately integrating near-infrared absorption and visible light reflection, and obtaining basic palmprint information data and basic palm vein detection data; establishing a palmprint and palm vein basic information database based on the basic palmprint information data and basic palm vein detection data, extracting palmprint information data features and palm vein data features, and constructing a palmprint and palm vein integrated recognition deep learning model; training a palmprint and palm vein integrated recognition deep learning model based on the palmprint and palm vein basic information database, and obtaining a palmprint and palm vein integrated recognition deep learning precision model; detecting palmprint information data and palm vein detection data in real time and inputting the palmprint and palm vein integrated recognition deep learning precision model, outputting the palmprint and palm vein integrated recognition result, and triggering the palmprint and palm vein integrated recognition when any palmprint and palm vein recognition is abnormal. Abnormal palm vein recognition alarm; can solve the problem of high-speed intelligent recognition of the rapid development of palm print and palm vein recognition data; can more accurately extract palm texture features, solve the problem of abnormal absorption rate of deoxyhemoglobin and other physiological tissues to near-infrared light caused by skin surface damage, and has significant improvement on the influence of factors such as skin aging; can detect palms in various states and accurately obtain basic palm print information data and basic palm vein detection data; can obtain accurate palm print information data features and palm vein data features and build a palm print and palm vein integrated recognition model, and the model accuracy is significantly improved; can train the model and make palm print and palm vein recognition more accurate; can detect palm print information data and palm vein detection data in real time and accurately output palm print and palm vein integrated recognition results, can promptly detect palm print and palm vein recognition abnormalities and give timely alarms, and significantly improve the real-time accuracy of palm print information and palm vein detection.

[0057] In one embodiment, S100 includes:

[0058] S101, adopting near-infrared absorption and visible light reflection dual-light alternating integrated detection, setting a near-infrared absorption dual-light alternating integrated detector;

[0059] S102, alternately irradiating the palm with five fingers stretched out with visible light and near-infrared rays of a near-infrared absorption dual-light alternating integrated detector, and obtaining basic palm print information data and basic palm vein detection data through visible light ray reflection sensing and near-infrared ray reflection sensing detection.

[0060] The principle and effect of the above technical solution are as follows: adopt near-infrared absorption and visible light reflection dual-light alternating integrated detection, set up a near-infrared absorption dual-light alternating integrated detector; alternately irradiate the palm when the five fingers are stretched out with visible light rays and near-infrared rays of the near-infrared absorption dual-light alternating integrated detector, respectively, and obtain basic palm print information data and basic palm vein detection data through visible light ray reflection sensing and near-infrared ray reflection sensing detection respectively; the near-infrared absorption dual-light alternating integrated detector includes: a near-infrared ray emitting unit, a near-infrared ray reflection sensing unit, a palm vein absorption detection unit, a visible light ray emitting unit, a visible light ray reflection sensing unit, and a palm print collection and detection unit; the near-infrared ray emitting unit emits a first near-infrared ray to the palm to be detected; the near-infrared ray reflection sensing unit senses the first near-infrared ray absorption reflection line after the near-infrared absorption of the palm vein to be detected; the palm vein absorption detection unit detects the first The near-infrared ray absorption reflection line is obtained to obtain first near-infrared ray absorption reflection line detection information; the visible light ray emitting unit emits the first visible light ray to the palm to be detected; the visible light ray reflection sensing unit senses the first visible light ray reflection line of the visible light reflected by the palm to be detected; the palm print collection and detection unit detects the first visible light ray reflection line to obtain first visible light ray reflection line detection information; the palm when the five fingers are stretched out is alternately irradiated with visible light rays and near-infrared rays to obtain second near-infrared ray absorption reflection line detection information and second visible light ray reflection line detection information; according to the first visible light ray reflection line detection information, the first near-infrared ray absorption reflection line detection information, the second near-infrared ray absorption reflection line detection information and the second visible light ray reflection line detection information, multiple groups of palm print information data and multiple groups of basic palm vein detection data are constructed to obtain basic palm print information data and basic palm vein detection data respectively.

[0061] In one embodiment, S200 includes:

[0062] S201, establishing a palm print and palm vein basic information database based on basic palm print information data and basic palm vein detection data;

[0063] S202, extracting palm print information data features and palm vein data features according to the palm print and palm vein basic information database;

[0064] S203, performing palm print and palm vein integrated identification through statistical analysis of palm print information data features and palm vein data features, and constructing a palm print and palm vein integrated identification deep learning model.

[0065] The principle and effect of the above technical solution are as follows: a palm print and palm vein basic information database is established based on basic palm print information data and basic palm vein detection data; palm print information data features and palm vein data features are extracted based on the palm print and palm vein basic information database; palm print and palm vein integrated judgment is performed through statistical analysis of palm print information data features and palm vein data features, and a palm print and palm vein integrated recognition deep learning model is constructed; the palm print and palm vein integrated recognition deep learning model includes: a palm print trunk feature extraction network, a palm vein trunk feature extraction network, an input layer, a palm print and palm vein weighted integration hidden layer, an output layer and a palm print and palm vein integrated comparison layer; a palm print trunk feature extraction network is constructed, and a palm print and palm vein integrated recognition deep learning model is constructed; a palm print and palm vein trunk feature extraction network is constructed, and ... The palm vein feature extraction network is used to extract palm vein features; the palm vein trunk feature extraction network extracts palm vein features; the input layer inputs the palm vein features and the palm print features into the palm print and palm vein weighted integration hidden layer respectively, and the palm print and palm vein weighted integration hidden layer performs weighted integration according to the input layer data respectively, and the output layer outputs the palm vein feature recognition information and the palm print feature recognition information after the weighted palm vein feature data and the weighted palm print feature data; the integrated comparison layer compares the palm print feature recognition information and the palm vein feature recognition information with the basic palm print information data and the basic palm vein detection data to determine whether they are the palm print and palm vein of the same individual; and outputs the palm print and palm vein integrated recognition judgment result.

[0066] In one embodiment, S300 includes:

[0067] S301, establishing a palm print and palm vein training set, a palm print and palm vein verification set, and a palm print and palm vein test set based on a palm print and palm vein basic information database, and training a palm print and palm vein integrated recognition deep learning model;

[0068] S302, obtaining a precise palm print and palm vein integrated recognition deep learning model by training the palm print and palm vein integrated recognition deep learning model until the output of the palm print and palm vein integrated recognition deep learning model reaches a preset data accuracy.

[0069] The principle and effect of the above technical solution are: according to the palm print and palm vein basic information database, a palm print and palm vein training set, a palm print and palm vein verification set and a palm print and palm vein test set are established to train a palm print and palm vein integrated recognition deep learning model; by training the palm print and palm vein integrated recognition deep learning model until the output of the palm print and palm vein integrated recognition deep learning model reaches the preset data accuracy, a palm print and palm vein integrated recognition deep learning precision model is obtained; working voltage DC12V; working temperature -25℃-65℃; recognition speed, single recognition ≤300ms; recognition accuracy, recognition rate ≥99.99%, misrecognition rate ≤0.002%.

[0070] In one embodiment, S400 includes:

[0071] S401, real-time detection of palm print information data and palm vein detection data, open the palm naturally, place the palm 5-10 cm above the sensor, and shake it slightly; obtain real-time detection of palm print information data and real-time detection of palm vein detection data;

[0072] S402, inputting the real-time palm print information data and the real-time palm vein detection data into the palm print and palm vein integrated recognition deep learning precision model;

[0073] S403, the palm print and palm vein integrated recognition deep learning precision model outputs the palm print and palm vein integrated recognition result;

[0074] S404, triggering a palm print and palm vein recognition abnormality alarm when any one of the palm print and palm vein recognition is abnormal according to the palm print and palm vein integrated recognition result.

[0075] The principle and effect of the above technical solution are: real-time detection of palm print information data and palm vein detection data, the palm is naturally opened, the palm is placed 5-10 cm above the sensor, and it is slightly swayed; real-time detection of palm print information data and real-time detection of palm vein detection data are obtained; the real-time detection of palm print information data and real-time detection of palm vein detection data are input into the palm print and palm vein integrated recognition deep learning precision model; the palm print and palm vein integrated recognition deep learning precision model outputs the palm print and palm vein integrated recognition result; according to the palm print and palm vein integrated recognition result, when any palm print and palm vein recognition is abnormal, the palm print and palm vein recognition abnormal alarm is triggered; the palm is naturally opened, the palm is placed 5-10 cm above the sensor, and it is slightly swayed.

[0076] The present invention provides a palmprint data recognition system based on deep learning, comprising:

[0077] The palm print and palm vein basic detection module detects the palm of the hand with five fingers stretched out by alternately absorbing near-infrared light and reflecting visible light, and obtains basic palm print information data and basic palm vein detection data;

[0078] The palm print and palm vein integrated recognition architecture module establishes a palm print and palm vein basic information database based on basic palm print information data and basic palm vein detection data, extracts palm print information data features and palm vein data features, and constructs a palm print and palm vein integrated recognition deep learning model;

[0079] Palm print and palm vein recognition model training module, based on the palm print and palm vein basic information database, trains a palm print and palm vein integrated recognition deep learning model to obtain a palm print and palm vein integrated recognition deep learning accurate model;

[0080] The real-time integrated palm print and palm vein recognition module detects palm print information data and palm vein detection data in real time and inputs them into the palm print and palm vein integrated recognition deep learning precision model, outputs the palm print and palm vein integrated recognition results, and triggers the palm print and palm vein recognition abnormality alarm when any of the palm print and palm vein recognition is abnormal.

[0081] The principle and effect of the above technical solution are as follows: the present invention provides a palm print data recognition system based on deep learning, including: a palm print and palm vein basic detection module 100, which detects the palm of the five fingers when they are stretched out by alternately absorbing near-infrared light and reflecting visible light, and obtains basic palm print information data and basic palm vein detection data; a palm print and palm vein integrated recognition architecture module 200, which establishes a palm print and palm vein basic information database based on the basic palm print information data and the basic palm vein detection data, extracts palm print information data features and palm vein data features, and constructs a palm print and palm vein integrated recognition deep learning model; a palm print and palm vein recognition model training module 300, which trains a palm print and palm vein integrated recognition deep learning model based on the palm print and palm vein basic information database, and obtains a palm print and palm vein integrated recognition deep learning precision model; a palm print and palm vein real-time integrated recognition module 400, which detects palm print information data and palm vein detection data in real time and inputs them into the palm print and palm vein integrated recognition deep learning precision model, and outputs It can output palm print and palm vein integrated recognition results, and trigger palm print and palm vein recognition abnormality alarm when any palm print and palm vein recognition is abnormal; it can solve the problem of high-speed intelligent recognition with rapid development of palm print and palm vein recognition data; it can extract palm texture features more accurately, solve the problem of abnormal absorption rate of deoxyhemoglobin and other physiological tissues to near-infrared light caused by skin surface damage, and has significant improvement on the influence of factors such as skin aging; it can detect palms in various states and accurately obtain basic palm print information data and basic palm vein detection data; it can obtain accurate palm print information data features and palm vein data features and construct a palm print and palm vein integrated recognition model, and the model accuracy is significantly improved; it can train the model and make palm print and palm vein recognition more accurate; it can detect palm print information data and palm vein detection data in real time and accurately output palm print and palm vein integrated recognition results, and can promptly detect palm print and palm vein recognition abnormalities and give timely alarms, and the real-time accuracy of palm print information and palm vein detection is significantly improved.

[0082] In one embodiment, the palm print and palm vein basic detection module includes:

[0083] The dual-light alternating integrated setting unit adopts near-infrared absorption and visible light reflection dual-light alternating integrated detection, and sets a near-infrared absorption dual-light alternating integrated detector;

[0084] The absorption and reflection dual-light alternating detection unit uses visible light and near-infrared rays of the near-infrared absorption dual-light alternating integrated detector to alternately irradiate the palm when the five fingers are stretched out, and obtains basic palm print information data and basic palm vein detection data through visible light reflection sensing and near-infrared reflection sensing detection.

[0085] The principle and effect of the above technical solution are as follows: the palm print and palm vein basic detection module includes: a dual-light alternating integrated setting unit, which adopts near-infrared absorption and visible light reflection dual-light alternating integrated detection, and sets a near-infrared absorption dual-light alternating integrated detector; the absorption and reflection dual-light alternating detection unit, which respectively irradiates the palm when the five fingers are stretched out with visible light rays and near-infrared rays of the near-infrared absorption dual-light alternating integrated detector, and obtains basic palm print information data and basic palm vein detection data through visible light ray reflection sensing and near-infrared ray reflection sensing detection; the near-infrared absorption dual-light alternating integrated detector includes: a near-infrared ray emitting unit 11, a near-infrared ray reflection sensing unit 12, a palm vein absorption detection unit 13, a visible light ray emitting unit 14, a visible light ray reflection sensing unit 15, and a palm print collection detection unit 16; the near-infrared ray emitting unit emits a first near-infrared ray to the palm to be detected; the near-infrared ray reflection sensing unit senses the palm vein to be detected after near-infrared absorption a first near-infrared ray absorption reflection line; a palm vein absorption detection unit detects the first near-infrared ray absorption reflection line and obtains first near-infrared ray absorption reflection line detection information; a visible light ray emitting unit emits a first visible light ray to the palm to be detected; a visible light ray reflection sensing unit senses the first visible light ray reflection line reflected by the visible light of the palm to be detected; a palm print collection and detection unit detects the first visible light ray reflection line and obtains first visible light ray reflection line detection information; by alternately irradiating the palm with five fingers stretched out with visible light rays and near-infrared rays, second near-infrared ray absorption reflection line detection information and second visible light ray reflection line detection information are obtained; according to the first visible light ray reflection line detection information, the first near-infrared ray absorption reflection line detection information, the second near-infrared ray absorption reflection line detection information and the second visible light ray reflection line detection information, multiple groups of palm print information data and multiple groups of basic palm vein detection data are constructed to obtain basic palm print information data and basic palm vein detection data respectively.

[0086] In one embodiment, the palm print and palm vein integrated recognition architecture module includes:

[0087] A palm print and palm vein basic information database unit is used to establish a palm print and palm vein basic information database based on basic palm print information data and basic palm vein detection data;

[0088] The information data feature extraction unit extracts the palm print information data features and palm vein data features according to the palm print and palm vein basic information database;

[0089] The integrated recognition deep learning model unit performs palm print and palm vein integrated judgment through statistical analysis of palm print information data features and palm vein data features, and constructs a palm print and palm vein integrated recognition deep learning model.

[0090] The principle and effect of the above technical solution are as follows: the palm print and palm vein integrated recognition architecture module includes: a palm print and palm vein basic information library unit, which establishes a palm print and palm vein basic information library based on basic palm print information data and basic palm vein detection data; an information data feature extraction unit, which extracts palm print information data features and palm vein data features based on the palm print and palm vein basic information library; an integrated recognition deep learning model unit, which performs palm print and palm vein integrated judgment through statistical analysis of palm print information data features and palm vein data features, and constructs a palm print and palm vein integrated recognition deep learning model; the palm print and palm vein integrated recognition deep learning model includes: a palm print trunk feature extraction network, a palm vein trunk feature extraction network, an input layer, a palm print and palm vein weighted set The invention is a method for extracting palm vein features from a palm print and palm vein by using a hidden layer, an output layer and a palm print and palm vein integrated comparison layer; the palm print trunk feature extraction network extracts palm print features; the palm vein trunk feature extraction network extracts palm vein features; the input layer inputs the palm vein features and the palm print features into the palm print and palm vein weighted integration hidden layer respectively, the palm print and palm vein weighted integration hidden layer performs weighted integration according to the input layer data respectively, the output layer outputs the palm vein feature recognition information and the palm print feature recognition information after the weighted palm vein feature data and the weighted palm print feature data; the integrated comparison layer compares the palm print feature recognition information and the palm vein feature recognition information with the basic palm print information data and the basic palm vein detection data to determine whether they are the palm print and palm vein of the same individual; and outputs the palm print and palm vein integrated recognition judgment result.

[0091] In one embodiment, the palm print and palm vein recognition model training module includes:

[0092] The deep learning model training unit establishes a palm print and palm vein training set, a palm print and palm vein verification set, and a palm print and palm vein test set based on the palm print and palm vein basic information database, and trains a palm print and palm vein integrated recognition deep learning model;

[0093] The deep learning precision model unit obtains the deep learning precision model of palm print and palm vein integrated recognition by training the deep learning model of palm print and palm vein integrated recognition until the output of the deep learning model of palm print and palm vein integrated recognition reaches the preset data accuracy.

[0094] The principle and effect of the above technical solution are: a palm print and palm vein recognition model training module, including: a deep learning model training unit, which establishes a palm print and palm vein training set, a palm print and palm vein verification set and a palm print and palm vein test set according to a palm print and palm vein basic information database, and trains a palm print and palm vein integrated recognition deep learning model; a deep learning precision model unit, which obtains a palm print and palm vein integrated recognition deep learning precision model by training the palm print and palm vein integrated recognition deep learning model until the output of the palm print and palm vein integrated recognition deep learning model reaches a preset data accuracy; working voltage DC12V; working temperature -25℃-65℃; recognition speed, single recognition ≤300ms; recognition accuracy, recognition rate ≥99.99%, misrecognition rate ≤0.002%.

[0095] In one embodiment, the palm print and palm vein real-time integrated recognition module includes:

[0096] Palm print and palm vein real-time detection unit, real-time detection of palm print information data and palm vein detection data, open your palm naturally, place your palm 5-10 cm above the sensor, and swing it slightly; obtain real-time detection of palm print information data and real-time detection of palm vein detection data;

[0097] A detection data input unit, which inputs real-time palm print information data and real-time palm vein detection data into a palm print and palm vein integrated recognition deep learning precision model;

[0098] Integrated recognition deep learning output unit: the palm print and palm vein integrated recognition deep learning precision model outputs the palm print and palm vein integrated recognition results;

[0099] The palm print and palm vein recognition abnormality alarm unit triggers a palm print and palm vein recognition abnormality alarm according to the palm print and palm vein integrated recognition result when any of the palm print and palm vein recognition is abnormal.

[0100] The principle and effect of the above technical solution are: a palm print and palm vein real-time integrated recognition module, including: a palm print and palm vein real-time detection unit, which detects palm print information data and palm vein detection data in real time, the palm is naturally opened, the palm is placed 5-10 cm above the sensor, and it is slightly swung; real-time detection palm print information data and real-time detection palm vein detection data are obtained; a detection data input unit, which inputs the real-time detection palm print information data and real-time detection palm vein detection data into a palm print and palm vein integrated recognition deep learning precision model; an integrated recognition deep learning output unit, the palm print and palm vein integrated recognition deep learning precision model outputs the palm print and palm vein integrated recognition result; a palm print and palm vein recognition abnormality alarm unit, according to the palm print and palm vein integrated recognition result, when any palm print and palm vein recognition is abnormal, triggers the palm print and palm vein recognition abnormality alarm; the palm is naturally opened, the palm is placed 5-10 cm above the sensor, and it is slightly swung.

[0101] Although the embodiments of the present invention have been disclosed as above, they are not limited to the applications listed in the specification and the implementation methods. They can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily implemented. Therefore, without departing from the general concept defined by the claims and the scope of equivalents, the present invention is not limited to the specific details and the illustrations shown and described herein.

Claims

1. A palmprint data recognition method based on deep learning, characterized in that: include: S100, detects the palm of the hand when five fingers are stretched out by alternately absorbing near-infrared light and reflecting visible light, and obtains basic palm print information data and basic palm vein detection data; S200, establishing a palm print and palm vein basic information database based on basic palm print information data and basic palm vein detection data, extracting palm print information data features and palm vein data features, and constructing a palm print and palm vein integrated recognition deep learning model; S300, training a palm print and palm vein integrated recognition deep learning model based on a palm print and palm vein basic information database, and obtaining a palm print and palm vein integrated recognition deep learning accurate model; S400, real-time detection of palm print information data and palm vein detection data and input into the palm print and palm vein integrated recognition deep learning precision model, output palm print and palm vein integrated recognition results, and trigger palm print and palm vein recognition abnormality alarm when any palm print and palm vein recognition is abnormal; S100 includes: S101, adopting near-infrared absorption and visible light reflection dual-light alternating integrated detection, setting a near-infrared absorption dual-light alternating integrated detector; S102, alternately irradiate the palm of the hand with five fingers stretched out with visible light and near infrared rays of the near infrared absorption dual light alternating integrated detector, and respectively obtain basic palm print information data and basic palm vein detection data through visible light reflection sensing and near infrared reflection sensing detection; the near infrared absorption dual light alternating integrated detector comprises: a near infrared ray emitting unit, a near infrared ray reflection sensing unit, a palm vein absorption detection unit, a visible light ray emitting unit, a visible light ray reflection sensing unit, and a palm print collection detection unit; the near infrared ray emitting unit emits a first near infrared ray to the palm to be detected; the near infrared ray reflection sensing unit senses a first near infrared ray absorption reflection line after the near infrared absorption of the palm vein to be detected; the palm vein absorption detection unit detects the first near infrared ray absorption reflection line and obtains the first near infrared ray absorption reflection line. Ray detection information; a visible light ray emitting unit emits a first visible light ray to a palm to be detected; a visible light ray reflection sensing unit senses a first visible light ray reflection line of visible light reflected from the palm to be detected; a palm print collection and detection unit detects the first visible light ray reflection line to obtain first visible light ray reflection line detection information; by alternately irradiating the palm with five fingers stretched out with visible light rays and near-infrared rays, second near-infrared ray absorption reflection line detection information and second visible light ray reflection line detection information are obtained; according to the first visible light ray reflection line detection information, the first near-infrared ray absorption reflection line detection information, the second near-infrared ray absorption reflection line detection information and the second visible light ray reflection line detection information, multiple groups of palm print information data and multiple groups of basic palm vein detection data are constructed to obtain basic palm print information data and basic palm vein detection data respectively; S200 includes: S201, establishing a palm print and palm vein basic information database based on basic palm print information data and basic palm vein detection data; S202, extracting palm print information data features and palm vein data features according to the palm print and palm vein basic information database; S203, through the statistical analysis of the palm print information data features and the palm vein data features, palm print and palm vein integrated judgment is performed to construct a palm print and palm vein integrated recognition deep learning model; the palm print and palm vein integrated recognition deep learning model includes: a palm print trunk feature extraction network, a palm vein trunk feature extraction network, an input layer, a palm print and palm vein weighted integration hidden layer, an output layer and a palm print and palm vein integrated comparison layer; the palm print trunk feature extraction network extracts palm print features; the palm vein trunk feature extraction network extracts palm vein features; the input layer inputs the palm vein features and the palm print features into the palm print and palm vein weighted integration hidden layer respectively, the palm print and palm vein weighted integration hidden layer performs weighted integration according to the input layer data respectively, and the output layer outputs the palm vein feature recognition information and the palm print feature recognition information of the weighted palm vein feature data and the weighted palm print feature data; the integrated comparison layer compares the palm print feature recognition information and the palm vein feature recognition information with the basic palm print information data and the basic palm vein detection data to determine whether they are the same individual palm print and palm vein; and outputs the palm print and palm vein integrated recognition judgment result.

2. A palmprint data recognition method based on deep learning according to claim 1, characterized in that: S300 includes: S301, establishing a palm print and palm vein training set, a palm print and palm vein verification set, and a palm print and palm vein test set based on a palm print and palm vein basic information database, and training a palm print and palm vein integrated recognition deep learning model; S302, obtaining a precise palm print and palm vein integrated recognition deep learning model by training the palm print and palm vein integrated recognition deep learning model until the output of the palm print and palm vein integrated recognition deep learning model reaches a preset data accuracy.

3. The palmprint data recognition method based on deep learning according to claim 1 is characterized in that: S400 includes: S401, real-time detection of palm print information data and palm vein detection data, open the palm naturally, place the palm 5-10 cm above the sensor, and shake it slightly; obtain real-time detection of palm print information data and real-time detection of palm vein detection data; S402, inputting the real-time palm print information data and the real-time palm vein detection data into the palm print and palm vein integrated recognition deep learning precision model; S403, the palm print and palm vein integrated recognition deep learning precision model outputs the palm print and palm vein integrated recognition result; S404, triggering a palm print and palm vein recognition abnormality alarm when any one of the palm print and palm vein recognition is abnormal according to the palm print and palm vein integrated recognition result.

4. A palmprint data recognition system based on deep learning, characterized in that: include: The palm print and palm vein basic detection module detects the palm of the hand with five fingers stretched out by alternately absorbing near-infrared light and reflecting visible light, and obtains basic palm print information data and basic palm vein detection data; The palm print and palm vein integrated recognition architecture module establishes a palm print and palm vein basic information database based on basic palm print information data and basic palm vein detection data, extracts palm print information data features and palm vein data features, and constructs a palm print and palm vein integrated recognition deep learning model; Palm print and palm vein recognition model training module, based on the palm print and palm vein basic information database, trains a palm print and palm vein integrated recognition deep learning model to obtain a palm print and palm vein integrated recognition deep learning accurate model; The palm print and palm vein real-time integrated recognition module detects palm print information data and palm vein detection data in real time and inputs them into the palm print and palm vein integrated recognition deep learning precision model, outputs the palm print and palm vein integrated recognition results, and triggers the palm print and palm vein recognition abnormality alarm when any palm print and palm vein recognition is abnormal; Palm print and palm vein basic detection module, including: The dual-light alternating integrated setting unit adopts near-infrared absorption and visible light reflection dual-light alternating integrated detection, and sets a near-infrared absorption dual-light alternating integrated detector; The absorption-reflection dual-light alternating detection unit uses the visible light rays and near-infrared rays of the near-infrared absorption dual-light alternating integrated detector to alternately illuminate the palm when the five fingers are stretched out, and obtains basic palm print information data and basic palm vein detection data through visible light ray reflection sensing and near-infrared ray reflection sensing detection respectively; The near-infrared absorption dual-light alternating integrated detector comprises: a near-infrared ray emitting unit, a near-infrared ray reflection sensing unit, a palm vein absorption detection unit, a visible light ray emitting unit, a visible light ray reflection sensing unit, and a palm print collection and detection unit; the near-infrared ray emitting unit emits a first near-infrared ray to the palm to be detected; the near-infrared ray reflection sensing unit senses a first near-infrared ray absorption reflection line after near-infrared absorption of the palm vein to be detected; the palm vein absorption detection unit detects the first near-infrared ray absorption reflection line and obtains detection information of the first near-infrared ray absorption reflection line; the visible light ray emitting unit emits a first visible light ray to the palm to be detected; the visible light ray reflection sensing unit senses the palm to be detected A first visible light ray reflection line reflected by visible light in the palm; a palm print collection and detection unit detects the first visible light ray reflection line to obtain first visible light ray reflection line detection information; by alternately irradiating the palm with five fingers stretched out with visible light rays and near-infrared rays, second near-infrared ray absorption reflection line detection information and second visible light ray reflection line detection information are obtained; according to the first visible light ray reflection line detection information, the first near-infrared ray absorption reflection line detection information, the second near-infrared ray absorption reflection line detection information and the second visible light ray reflection line detection information, multiple groups of palm print information data and multiple groups of basic palm vein detection data are constructed to obtain basic palm print information data and basic palm vein detection data respectively; Palm print and palm vein integrated recognition architecture module, including: A palm print and palm vein basic information database unit is used to establish a palm print and palm vein basic information database based on basic palm print information data and basic palm vein detection data; The information data feature extraction unit extracts the palm print information data features and palm vein data features according to the palm print and palm vein basic information database; The integrated recognition deep learning model unit performs palm print and palm vein integrated judgment through statistical analysis of palm print information data features and palm vein data features, and constructs a palm print and palm vein integrated recognition deep learning model; the palm print and palm vein integrated recognition deep learning model includes: a palm print trunk feature extraction network, a palm vein trunk feature extraction network, an input layer, a palm print and palm vein weighted integration hidden layer, an output layer and a palm print and palm vein integrated comparison layer; the palm print trunk feature extraction network extracts palm print features; the palm vein trunk feature extraction network extracts palm vein features; the input layer inputs the palm vein features and the palm print features into the palm print and palm vein weighted integration hidden layer respectively, the palm print and palm vein weighted integration hidden layer performs weighted integration according to the input layer data respectively, and the output layer outputs the palm vein feature recognition information and the palm print feature recognition information of the weighted palm vein feature data and the weighted palm print feature data; the integrated comparison layer compares the palm print feature recognition information and the palm vein feature recognition information with the basic palm print information data and the basic palm vein detection data to determine whether they are the same individual palm print and palm vein; and outputs the palm print and palm vein integrated recognition judgment result.

5. A palmprint data recognition system based on deep learning according to claim 4, characterized in that: Palm print and palm vein recognition model training module, including: The deep learning model training unit establishes a palm print and palm vein training set, a palm print and palm vein verification set, and a palm print and palm vein test set based on the palm print and palm vein basic information database, and trains a palm print and palm vein integrated recognition deep learning model; The deep learning precision model unit obtains the deep learning precision model of palm print and palm vein integrated recognition by training the deep learning model of palm print and palm vein integrated recognition until the output of the deep learning model of palm print and palm vein integrated recognition reaches the preset data accuracy.

6. A palmprint data recognition system based on deep learning according to claim 4, characterized in that: Palm print and palm vein real-time integrated recognition module, including: Palm print and palm vein real-time detection unit, real-time detection of palm print information data and palm vein detection data, open your palm naturally, place your palm 5-10 cm above the sensor, and swing it slightly; obtain real-time detection of palm print information data and real-time detection of palm vein detection data; A detection data input unit, which inputs real-time palm print information data and real-time palm vein detection data into a palm print and palm vein integrated recognition deep learning precision model; Integrated recognition deep learning output unit: the palm print and palm vein integrated recognition deep learning precision model outputs the palm print and palm vein integrated recognition results; The palm print and palm vein recognition abnormality alarm unit triggers a palm print and palm vein recognition abnormality alarm according to the palm print and palm vein integrated recognition result when any of the palm print and palm vein recognition is abnormal.

Images