Intelligent lock control method, device and electronic device based on image recognition

Abstract

The present application discloses an intelligent lock control method, device and electronic device based on image recognition. Among them, the intelligent lock control method based on image recognition includes: obtaining a first mode image through a first recognition module, performing image preprocessing on the first mode image to obtain a first preprocessed image; if the first preprocessed image is an image with unclear features, analyzing the unclear reasons for the image with unclear features; if the result of the reason analysis is that the lens is blocked, recording the occlusion position information corresponding to the occlusion, and confirming the camera type corresponding to the occlusion position information; if the first recognition module is a face module and the camera type is a near-infrared camera, sending a switching recognition module instruction; performing feature recognition on the third mode features, and confirming the opening of the intelligent lock according to the feature recognition result. This method enhances the anti-interference ability of the system while improving the recognition effectiveness and timeliness of biometric features, and maintains the recognition stability of the system.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition, and particularly to an intelligent lock control method, device and electronic device based on image recognition. Background Art

[0002] Today, with the increasing popularity of smart homes, intelligent door locks, as the first line of defense for access security, have attracted more and more attention. Among the authentication methods of intelligent door locks, those with relatively high security levels include face recognition and palm vein recognition, etc. Among them, 3D face recognition technology collects rich face information, such as facial shape, texture, depth, etc., and has higher accuracy, so as to compare with pre-stored face information; palm vein recognition technology uses near-infrared light to irradiate the palm, and the sensor senses the light reflected by the palm. The key lies in that hemoglobin in red blood cells flowing through veins will absorb near-infrared light near a wavelength of 760 nm, resulting in less reflection in the vein part. The position of the vein is identified by using the strength of the reflected near-infrared light to generate a vein pattern on the image. Then, the vein data read is compared with the pre-stored palm vein data to perform identification.

[0003] In a 3D face recognition system or a palm vein recognition system, generally, a camera module composed of a group of cameras and near-infrared lamps can be used to simultaneously realize face or palm vein recognition. The camera module is generally placed in the transparent panel area on the lock body for protection. However, it is inevitable that water droplets or dirt will adhere to the transparent panel area, thus affecting the shooting effect of the camera module and interfering with the accuracy and effectiveness of face or palm vein recognition. Summary of the Invention

[0004] Embodiments of the present invention provide an intelligent lock control method, device and electronic device based on image recognition to solve the problem that the camera module is generally set in the transparent panel area on the lock body for protection, but it is inevitable that water droplets or dirt will adhere to the transparent panel area, thus affecting the shooting effect of the camera module and interfering with the accuracy and effectiveness of face or palm vein recognition.

[0005] An intelligent lock control method based on image recognition includes:

[0006] Obtaining a first mode image through a first recognition module, and performing image preprocessing on the first mode image to obtain a first preprocessed image, where the first recognition module is any one of a face module or a palm vein module, and the second recognition module is the other module different from the first recognition module in the face module or the palm vein module;

[0007] If the first preprocessed image is an image with unclear features, then perform an analysis of the unclear reason on the image with unclear features to obtain a result of the reason analysis;

[0008] If the result of the cause analysis indicates that the lens is blocked, record the occlusion position information corresponding to the occlusion, confirm the camera type corresponding to the occlusion position information, and thus confirm whether the second recognition module is affected;

[0009] If the first recognition module is a face module and the camera type is a near-infrared camera, send a recognition module switching instruction to receive the third mode features input by the user through any one of the other recognition modules as the third recognition module;

[0010] Perform feature recognition on the third mode features, obtain the feature recognition result, and confirm the opening of the intelligent lock according to the feature recognition result.

[0011] An intelligent lock control device based on image recognition, comprising:

[0012] A first preprocessing image acquisition module, configured to acquire a first mode image through a first recognition module, perform image preprocessing on the first mode image, and acquire a first preprocessed image, wherein the first recognition module is any one of a face module or a palm vein module, and the second recognition module is the other module different from the first recognition module among the face module or the palm vein module;

[0013] A cause analysis result acquisition module, configured to perform unclear cause analysis on the unclear feature image if the first preprocessed image is an unclear feature image, and acquire the cause analysis result;

[0014] A camera type confirmation module, configured to record the occlusion position information corresponding to the occlusion, confirm the camera type corresponding to the occlusion position information, and thus confirm whether the second recognition module is affected if the result of the cause analysis indicates that the lens is blocked;

[0015] An instruction sending module, configured to send a recognition module switching instruction if the first recognition module is a face module and the camera type is a near-infrared camera, so as to receive the third mode features input by the user through any one of the other recognition modules as the third recognition module;

[0016] A feature recognition result acquisition module, configured to perform feature recognition on the third mode features, acquire the feature recognition result, and confirm the opening of the intelligent lock according to the feature recognition result.

[0017] An electronic device, comprising: a microprocessor supporting at least two biometric recognitions, and a functional module electrically connected to the microprocessor, the functional module including: a camera module, an infrared illumination light source, a face module, a palm vein module, a distance sensing module, a light sensing module, a memory module, an Ethernet communication module, a speaker, an electromagnetic lock control module, and a power supply module, and the Ethernet communication module is connected to a remote control end through Ethernet;

[0018] A memory module for storing preset feature data of at least two biometric identifications and a computer program that is in the memory module and can run on a microprocessor. When the computer program is executed by the microprocessor, the intelligent lock control method based on image recognition as described above is implemented; the microprocessor controls a camera module and a near-infrared light emitting tube to identify a face or a palm vein.

[0019] A light-sensitive module is used to detect ambient light information and send the ambient light information to the microprocessor, so that the microprocessor can switch the working mode of the camera in the camera module according to the ambient light information;

[0020] The camera module is arranged under the transparent panel and includes at least one visible light camera and at least one near-infrared camera. A visible light filter is embedded in the transparent panel; at least two near-infrared light emitting tubes are evenly distributed around each camera, and at least one near-infrared light emitting tube is evenly distributed between two cameras. The near-infrared light emitting tubes are used to provide infrared fill light function for the cameras in the camera module;

[0021] The intelligent lock control method, device and electronic device based on image recognition as described above, by analyzing the reasons for the biometric images that do not meet the recognition requirements obtained during shooting in a timely manner, can be converted into other effective biometric recognition modes in time when it is determined that the camera is blocked, such as dirty, enhancing the anti-interference ability of the system while improving the recognition effectiveness and timeliness of biometric features, and maintaining the recognition stability of the system. Description of the Drawings

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required to be used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0023] Figure 1 Illustrate a schematic diagram of the application environment of the intelligent lock control method based on image recognition in an embodiment of the present invention;

[0024] Figure 2 Illustrate the first flowchart of the intelligent lock control method based on image recognition in the first embodiment of the present invention;

[0025] Figure 3 Illustrate a schematic diagram of the intelligent lock of the intelligent lock control method based on image recognition in the first embodiment of the present invention;

[0026] Figure 4 Illustrate the second flowchart of the intelligent lock control method based on image recognition in an embodiment of the present invention;

[0027] Figure 5 Schematic diagram of a first preprocessed image with dirt taken by a camera in the intelligent lock control method based on image recognition in the first embodiment of the present invention;

[0028] Figure 6 Schematic diagram of a third flowchart of the intelligent lock control method based on image recognition in the fourth embodiment of the present invention;

[0029] Figure 7 Schematic diagram of a fourth flowchart of the intelligent lock control method based on image recognition in the seventh embodiment of the present invention;

[0030] Figure 8 Schematic diagram of the intelligent lock control device based on image recognition in the eighth embodiment of the present invention;

[0031] Figure 9 Schematic diagram of an electronic device in an embodiment of the present invention.

[0032] Explanation of reference numerals:

[0033] 10, visible light camera; 20, near-infrared camera; 30, infrared illumination light source. Detailed implementation manners

[0034] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0035] The intelligent lock control method based on image recognition provided by the embodiments of the present invention can be applied in an application environment such as Figure 1 . This intelligent lock control method based on image recognition is applied in an intelligent lock control system based on image recognition. The intelligent lock control system based on image recognition includes an intelligent lock and a remote control terminal. Among them, the intelligent lock communicates with the remote control terminal through a network, including various forms of intelligent terminals, such as a mobile phone terminal or a smart watch, etc., for realizing remote control unlocking of the intelligent lock. The intelligent lock determines whether the characteristics of the user match the preset biometric characteristics by performing biometric recognition on the user, and then controls the opening and closing of the door lock through an electromagnetic lock control module. This intelligent lock can be mainly applied in a smart home environment, and is particularly suitable for various places that require access control, such as residences, offices, hotels, etc., and can significantly improve the recognition accuracy of the intelligent lock in various environments. The remote control terminal can be realized by an independent remote control terminal, such as a mobile terminal or a remote control terminal cluster composed of multiple remote control terminals.

[0036] Concept explanation:

[0037] 3D Face Recognition: 3D structured light recognition uses a 3D camera module to generate a three-dimensional image, which can identify the three-dimensional coordinate information of each point in the field of view. As a result, the computer can obtain 3D data of the space and restore the complete three-dimensional world, and achieve various intelligent three-dimensional positioning. Among them, the 3D camera module includes a built-in speckle projector, a visible light camera (RGB camera) for obtaining 2D face information, a near-infrared camera for global exposure, and an infrared illumination source. The principle of 3D face recognition is to use a near-infrared projector to emit approximately 30,000 infrared structured light points onto the object being photographed. The wavelength of the near-infrared light used is 700-2526 nm, making the brightness of a specific illumination area or a specific visual target much higher than that of other targets and the surrounding area. The near-infrared camera captures the information to obtain the depth information of the face, and then through calculation, the 2D face information and the face depth information are superimposed and further processed to become 3D face information.

[0038] Palm Vein Recognition: Palm vein recognition utilizes the characteristic of hemoglobin in the blood to absorb near-infrared light. A near-infrared camera is used to obtain vein images under near-infrared light irradiation. Light with a wavelength between 750 nm and 1000 nm can penetrate the epidermis of the human body relatively well. The oxyhemoglobin and reduced hemoglobin in the blood have a relatively high absorption rate for near-infrared light in the wavelength range of 750 nm to 1000 nm, while the water in human tissues has a relatively low absorption rate in this wavelength range. Therefore, under near-infrared light irradiation, the palm vein part will appear dark due to absorbing more light, while other tissue cells in the palm will appear bright due to the reflection of water, thus obtaining a clear palm vein image with unique personal characteristics. Palm vein blood vessels grow under the epidermis of the human body and cannot be directly observed under visible light conditions.

[0039] Embodiment 1

[0040] As Figure 2 shown, a smart lock control method based on image recognition is provided. Taking the application of this method in the microprocessor of the smart lock in Figure 1 as an example, the specific steps are as follows:

[0041] S110. Obtain a first-mode image through the first recognition module, and perform image preprocessing on the first-mode image to obtain a first preprocessed image. Among them, the first recognition module is any one of the face module or the palm vein module, and the second recognition module is the other module different from the first recognition module in the face module or the palm vein module.

[0042] Specifically, the face module is a module for implementing face feature recognition, which may include a visible light camera 10, a near-infrared camera 20, and an infrared illumination light source 30; the palm vein module is a module for implementing palm vein feature recognition, including a near-infrared camera 20 and an infrared illumination light source 30, and the infrared illumination light source 30 may specifically include a near-infrared light emitting tube. At least two near-infrared light emitting tubes are distributed near each camera, and at least one near-infrared light emitting tube is evenly distributed between two cameras. The near-infrared light emitting tubes are used to provide infrared fill light function for the cameras.

[0043] Since both the face module and the palm vein module include a near-infrared camera 20 and an infrared illumination light source 30, and the face module has an additional visible light camera 10 compared to the palm vein module. To save production costs, most manufacturers generally set up a shared near-infrared camera 20 and an infrared illumination light source 30 when implementing the two recognition modes, as Figure 3 shown, and additionally set up a visible light camera 10 for face recognition. Through such a setting by the manufacturer, dual recognition of the face mode and the palm vein mode can be achieved through a relatively simple and low-cost architecture.

[0044] The difference between the two is that the effective distance of face recognition is mainly affected by the camera resolution, algorithm optimization level, and specific application scenarios. In a 1:1 scenario at a short distance (such as within 0.4 meters to 1 meter, especially 0.4 - 0.8 meters in front of the lens), the accuracy of face recognition is high. However, as the recognition distance increases, decreases, and the size of the face database increases, the accuracy of face recognition will gradually decline; while the effective distance of palm vein recognition is generally between 5 - 20 centimeters. This recognition method is not only convenient and easy to use, but also has a high level of security and hygiene. Palm vein recognition verifies the identity by capturing the vein images on the palm surface. Since the palm vein features of each person are unique, this recognition method has high accuracy and reliability.

[0045] Therefore, when using a shared camera to distinguish whether the biometric feature is a face or a palm vein, the distance determination method or the image graphic feature confirmation method can be used for determination. Currently, a distance sensor is generally used to confirm the features of the target biometric feature, and then the camera is used to capture the biometric feature image of the target biometric feature, which may be a face feature image or a palm vein feature image.

[0046] When the face module is used as the first recognition module, the palm vein module is used as the second recognition module; correspondingly, when the palm vein module is used as the first recognition module, the face recognition module is used as the second recognition module.

[0047] The first mode image is a biometric feature image obtained by capturing through the camera of the first recognition module, which may be a face feature image or a palm vein feature image.

[0048] The first preprocessed image is to remove the interfering information in the color image captured by the camera, reduce the amount of calculation. Before extracting eigenvalue such as correlation coefficient, edge information, and frequency domain energy, the image is grayscaled and denoised to highlight useful information, forming a preliminary feature image, which facilitates the subsequent analysis and calculation of the feature image, achieving a faster and more accurate purpose.

[0049] Furthermore, to extract features from the first mode image, it is first necessary to determine whether the image meets the feature significance standard. It can be understood that if the captured biometric features are interfered by environmental factors or hardware factors, etc., making it difficult to extract sufficient feature information that meets the matching requirements from the captured image, that is, when the first mode image does not meet the feature significance standard, the first preprocessed image becomes an image with unclear features and cannot continue with feature matching. At this time, the user should be reminded to switch the biometric input mode in a timely manner.

[0050] Furthermore, environmental factors include dim light, etc., and hardware factors include defects in the camera or occlusion on the transparent panel corresponding to the camera. The factors causing occlusion include water droplets, dirt, or oil stains, etc.

[0051] Preferably, for the intelligent lock control method based on image recognition provided in this embodiment, after step S110, as Figure 4 shown, that is, after obtaining the first preprocessed image, it further specifically includes the following steps:

[0052] S1101. If the first preprocessed image is an image with clear features, perform feature recognition based on the image with clear features to obtain a feature recognition result, and confirm the opening of the intelligent lock according to the feature recognition result.

[0053] Specifically, an image with clear features means that the first mode image captured by the system meets the requirements for feature extraction in terms of clarity and lighting, etc., and a feature recognition result that meets the comparison requirements can be extracted from the first mode image.

[0054] It can be understood that when the first preprocessed image is an image with clear features, the comparison of face or palm vein features can be smoothly carried out, and the opening of the intelligent lock can be quickly confirmed through the feature recognition result.

[0055] S120. If the first preprocessed image is an image with unclear features, perform an analysis of the unclear reason on the image with unclear features to obtain a reason analysis result.

[0056] Specifically, the feature-unclear image is an image in which the biometric information carried by the first mode image does not meet the feature significance standard. At this time, the present embodiment can continue to perform the unknown cause analysis based on the feature-unclear image, that is, to confirm what factors cause the feature to be unknown, and the analysis result, such as lens occlusion or dim light, is the cause analysis result.

[0057] The image recognition-based smart lock control method provided in this embodiment can continue to perform unknown cause analysis on the image with unknown features when the first preprocessed image is an image with unknown features, instead of immediately reminding the user to switch the recognition mode when the image is an image with unknown features. By obtaining the cause analysis results, the user can be more efficiently reminded which recognition mode to switch to, avoiding the use of invalid feature recognition modes again.

[0058] S130. If the cause analysis result is that the lens is blocked, the blocking position information corresponding to the blocking is recorded, and the camera type corresponding to the blocking position information is confirmed, so as to determine whether it affects the second recognition module.

[0059] Specifically, the face module and palm vein module provided in this embodiment share a near-infrared camera and an infrared illumination source. When there is dirt on the near-infrared camera, both modules cannot extract features. In the first pre-processed image with dirt, a lot of information is difficult to recognize, and although the area of mud on the transparent panel is very small, in the first pre-processed image captured by the camera, the area covered by dirt such as mud occupies a large part of the image area, seriously affecting the image clarity, such as Figure 5 shown.

[0060] If only the visible light camera of the face module is dirty, it will not affect the face module recognition. If the face module recognition fails, the system can prompt the user to continue the identity authentication in palm vein recognition mode, or the user can switch to other authentication methods according to their own preferences, such as fingerprint recognition, palm print recognition, password input or NFC card authentication.

[0061] This embodiment can use an image edge information algorithm to extract the pixel position of the obstruction such as water droplets or dirt on the image, that is, the obstruction position information. Image edge information is very important information of the image. It is generally believed that the edge of the object in the image is the area where the gray value changes. The change of the gray value in the tangential direction of the edge is relatively gentle, and the change of the gray value in the normal direction of the edge is relatively drastic. At the same time, the edge information of the image can also represent the change of the image depth.

[0062] The user stands in front of the camera module. The depth of field is relatively large in the direct front of the camera module, and the depth information is relatively rich. There is more edge information in the area where the depth changes. Therefore, the image quality can be evaluated using the edge information of the image. When the lens surface is clean, the image captured by the camera is relatively clear and sharp, and the image contains more edge information. When the lens surface is covered with dirt, it is difficult to distinguish objects from the image, no edge information can be detected in the dirt-covered area, and the edge information of the entire image is also less.

[0063] In this embodiment, a convolution operator can be used to perform convolution operations on the image for image edge extraction. The first-order operator can be one of the Robert operator, Prewitt operator, and Sobel operator. Convolution operations are performed on the image sequences captured by all cameras (including images without dirt and images with dirt), and the edge information in the image sequences of images without dirt and the image sequences of dirty images can be extracted to obtain the image edge information image. The edge information contained in the image sequence of images without dirt is more than that in the image sequence of dirty images. The edge information in the image can be used as a basis for judging whether the lens surface is dirty. If it is detected that the edge information of the image has been at a relatively low value or suddenly decreases, it is very likely that the camera surface is contaminated.

[0064] An image with more edge information in the image edge information image and obvious edges of the human face or palm vein is identified as a clear image, that is, an image without dirt.

[0065] The positions of a little image edges with less edge information in the image, especially in the area covered by dirt where almost no edge information can be extracted, are recorded as occlusion position information.

[0066] Based on this position information, the target camera corresponding to the captured first-mode image can be deduced, and it can be confirmed whether the type of the target camera is a near-infrared camera or a visible-light camera.

[0067] S140. If the first recognition module is a face module and the camera type is a near-infrared camera, a switching recognition module instruction is sent to receive the third-mode feature input by the user through any one of the other recognition modules as the third recognition module.

[0068] Specifically, in this embodiment, when the first recognition module is a face module and the camera type is a near-infrared camera, at this time, the second recognition module is a palm vein module. As can be seen from the foregoing, when the shared near-infrared camera is blocked, the imaging of both the face module and the palm vein module is affected. At this time, this embodiment can use a prompt method. During the day, it can be a combined prompt of text on the display screen interface and voice. At night, it can only be a text prompt on the interface of the smart lock to prompt the user to switch to a third mode feature other than face recognition and palm vein recognition modes. The third mode feature can be confirmed according to the user's current input, such as fingerprint mode feature, palmprint mode feature, password input, NFC wireless signal, Bluetooth wireless signal, or terminal remote control signal lamp.

[0069] The intelligent lock control method based on image recognition provided by this embodiment can timely remind the user to switch to an effective third mode feature, avoiding the user from repeatedly inputting unrecognizable features and thus affecting the recognition efficiency.

[0070] S150. Perform feature recognition on the third mode feature, obtain the feature recognition result, and confirm the opening of the intelligent lock according to the feature recognition result.

[0071] Specifically, this embodiment can perform feature recognition according to the third mode feature input by the user to obtain the feature recognition result. When the input third mode feature matches the preset third biometric feature stored in the intelligent lock, it can be confirmed that the user is an authorized user, and an opening signal can be sent to the electromagnetic lock control module of the intelligent lock to open the intelligent lock.

[0072] The intelligent lock control method based on image recognition provided by this embodiment can timely analyze the reason for the biometric feature image that does not meet the recognition requirements obtained during shooting. When it is determined that the camera is blocked, such as being dirty, it can be timely switched to other effective biometric recognition modes, enhancing the anti-interference ability of the system while improving the recognition effectiveness and timeliness of biometric features, and maintaining the recognition stability of the system.

[0073] Embodiment 2

[0074] The intelligent lock control method based on image recognition provided by this embodiment, after step S130, as Figure 4 shown, that is, after confirming the camera type corresponding to the occlusion position information, further includes the following steps:

[0075] S1301. If the first recognition module is a face module and the camera type is a visible light camera, send a recognition module switching instruction to receive the third mode feature input by the user through any one of the other recognition modules as the third recognition module, or receive the second mode feature input by the user through the second recognition module.

[0076] S1302. If the first recognition module is a palm vein module and the camera type is a near-infrared camera, send a recognition module switching instruction to receive third-mode features input by the user using any one of the other recognition modules as the third recognition module.

[0077] S1303. Perform feature recognition on the third-mode features or the second-mode features, obtain a feature recognition result, and confirm the opening of the smart lock according to the feature recognition result.

[0078] Specifically, when the first recognition module is a face recognition module and the camera type is a visible light camera, that is, at this time the occlusion exists in the visible light camera, which has an impact on face recognition module recognition but has no impact on palm vein recognition. At this time, the palm vein module as the second recognition module can continue to perform pattern recognition. In this embodiment, through a prompting method, during the day, it can be a combined prompt of text on the display screen interface and voice, and at night, it can only be a text prompt on the smart lock interface to prompt the user to switch to the second-mode features and third-mode features other than face recognition. The third-mode features can be confirmed according to the user's current input, such as fingerprint mode features, palmprint mode features, password input, NFC wireless signal, Bluetooth wireless signal, or terminal remote control signal lamp.

[0079] When the first recognition module is a palm vein module and the camera type is a near-infrared camera, that is, at this time the occlusion exists in the near-infrared camera, which has an impact on both face recognition module recognition and palm vein recognition. At this time, the face recognition module as the second recognition module cannot continue to perform pattern recognition either. In this embodiment, through a prompting method, prompt the user to switch to the third-mode features other than face recognition and palm vein recognition, perform feature extraction and feature comparison, so as to control the opening of the smart lock.

[0080] There is no situation where the first recognition module is a palm vein module and the camera type is a visible light camera. Since the palm vein module only activates the near-infrared camera, this situation does not need to be discussed.

[0081] The implementation process of 3D face recognition generally requires four steps: face image acquisition and detection, face image preprocessing, face image feature extraction, and face image matching and recognition.

[0082] 1. Face image acquisition and detection: Record as a pattern image through methods such as a camera in the form of static images, dynamic images, etc., and determine the position and size of the face through face detection. Specifically, the Adaboost face detection algorithm can be used, based on integral images, cascade detectors, and the Adaboost algorithm, etc., to detect the position and image information of the frontal face.

[0083] 2. Facial Image Preprocessing: Due to factors such as the lighting at the image acquisition location being different, there are significant differences in the quality of the images. Image preprocessing can remove the information that interferes with the image, improve the quality of the picture, highlight the useful information as features to clarify the image, and facilitate the subsequent analysis and calculation of the feature image.

[0084] 3. Facial Image Feature Extraction: By analyzing and calculating the feature - clarified image after preprocessing, extract the facial features of the portrait in the image, such as the characteristics of the nose, eyes, chin, mouth, etc. Then calculate the positional relationships between these features and use their geometric features as the feature information to be compared for face recognition.

[0085] 4. Facial Image Matching and Recognition: Search and compare the feature information to be compared with the user image information stored in the own database. When the matching degree of the two reaches the matching threshold, the user information can be confirmed to achieve the purpose of unlocking.

[0086] The implementation process of palm vein recognition generally also requires four steps: palm vein image acquisition and detection, palm vein image preprocessing, palm vein image feature extraction, and palm vein image matching and recognition.

[0087] 1. Palm Vein Image Acquisition and Detection: Record as a pattern image through methods such as cameras in the forms of static images, dynamic images, etc. When obtaining the palm vein image, the back of the hand does not need to contact the device, and it can be easily placed to complete the recognition. In this way, the hand does not need to contact the device, and it will not be unable to recognize due to dirt on the hand. Moreover, the influence of external factors such as temperature can be ignored.

[0088] 2. Palm Vein Image Preprocessing: Due to factors such as the lighting at the image acquisition location being different, there are significant differences in the quality of the images. The main purpose of image preprocessing is to remove the information that interferes with the image, improve the quality of the picture, and highlight the useful information as features to clarify the image. Preprocessing includes vein segmentation, smoothing and thinning, vein thinning, and burr trimming, etc., to obtain a feature - clarified image that meets the comparison standard.

[0089] 3. Palm Vein Image Feature Extraction: By analyzing and calculating the feature - clarified image after preprocessing, extract the feature information to be compared of the palm vein in the image, such as the moment feature vector of the vein skeleton, etc.

[0090] 4. Palm Vein Image Matching and Recognition: Adopt a feature comparison algorithm, such as a support vector machine, etc., to search and compare the feature information to be compared with the user image information stored in the own database. When the matching degree of the two reaches a certain ratio, the user information can be confirmed to achieve the purpose of unlocking.

[0091] Example Three

[0092] The intelligent lock control method based on image recognition provided in this embodiment, after step S140, that is, after sending the switching recognition module instruction, further specifically includes the following steps:

[0093] S1401. Send a panel cleaning instruction to clean the transparent panel corresponding to the near-infrared camera.

[0094] Specifically, the panel cleaning instruction is an instruction to notify the system or the user to clean the lens. If the system is equipped with a self-cleaning mechanism, the self-cleaning operation can be started based on the panel cleaning instruction at this time. If the system uses a prompting method, during the day, it can be a combined text and voice prompt on the display screen interface, and at night, it can be only a text prompt on the intelligent lock interface to prompt the user to clean the specific target lens with occlusion and dirt, so as to restore the normal use functions of the face recognition mode and palm vein recognition mode of the intelligent lock.

[0095] S1402. After a preset cleaning waiting time, determine the information change result of the occlusion position information.

[0096] Specifically, machine cleaning or manual cleaning takes a certain amount of time. In this application, after waiting for the time used to clean the camera, that is, after the preset cleaning waiting time, an image can be captured of the target lens that was previously dirty to confirm whether the dirt at the occlusion has been effectively cleaned, that is, whether the image at the occlusion position has changed. The information change result includes no change in position and a reduction in position information. The reduction in position information can be less than the original occlusion position information or there is no occlusion position information. A reduction in the position information less than the original indicates that the dirt on the target lens has not been completely cleaned. At this time, if a biometric feature can be captured, it does not affect the face or palm vein recognition mode. Specifically, it still depends on whether the first mode feature captured by the lens matches the preset first feature.

[0097] S1403. If the information change result is a reduction in position information, repeat the step of obtaining the first mode image by the first recognition module.

[0098] Specifically, the determination process of the information change result includes: obtaining the occlusion position information corresponding to the current image captured by the target camera corresponding to the dirt as the second position information. Taking the occlusion position information corresponding to the image with unclear features as the first position information, and comparing whether the second position information is less than the first position information. When the second position information is less than the first position information, it indicates that the area of the contaminated area has shrunk.

[0099] S1404. If the information change result is no change in position information, send a repair request message to the remote control end, and the repair request message includes the occlusion position information.

[0100] Specifically, when the result of information change indicates that the position information remains unchanged, it means that the dirt has not been effectively removed. It may be that there is a problem with the internal hardware of the camera at the lens. At this time, a repair message needs to be sent to the remote control terminal so that the maintenance personnel can subsequently confirm the specific location of the hardware damage according to the occlusion position information in a timely manner, improving the maintenance efficiency.

[0101] Embodiment 4

[0102] For the intelligent lock control method based on image recognition provided in this embodiment, after step S120, that is, after obtaining the cause analysis result, as Figure 4 and Figure 6 shown, it further specifically includes the following steps:

[0103] S1201. If the cause analysis result is that the light of the lens is insufficient, record the light source position information corresponding to the light source with insufficient light, confirm the infrared light source position corresponding to the light source position information, so as to confirm whether it affects the second recognition module.

[0104] S1202. If the first recognition module is a face module and the infrared light source position is the face infrared light source, send a recognition module switching instruction to receive the third mode features input by the user through any one of the other recognition modules as the third recognition module, or receive the second mode features input by the user through the second recognition module.

[0105] S1203. If the first recognition module is a face module and the infrared light source position is the common infrared light source, send a recognition module switching instruction to receive the third mode features input by the user through any one of the other recognition modules as the third recognition module.

[0106] S1204. If the first recognition module is a palm vein module and the infrared light source position is the common infrared light source, send a recognition module switching instruction to receive the third mode features input by the user through any one of the other recognition modules as the third recognition module.

[0107] S1205. If the first recognition module is a face module and the infrared light source position is the palm vein infrared light source, send a recognition module switching instruction to receive the third mode features input by the user through any one of the other recognition modules as the third recognition module, or receive the second mode features input by the user through the second recognition module.

[0108] S1206. Perform feature recognition on the third mode features or the second mode features, obtain the feature recognition result, and confirm the opening of the intelligent lock according to the feature recognition result.

[0109] Specifically, when the light is insufficient, the system can use multiple infrared illumination light sources, namely near-infrared emitting diodes, which are set to improve the shooting brightness. The near-infrared emitting diodes are precise active light sources, and after long-term use, they may be damaged and attenuated, which will also affect the normal shooting function of the camera.

[0110] When the camera is in the near-infrared shooting mode corresponding to the face module, the wavelength of the near-infrared rays used is 700 - 2526 nm, preferably 700 - 1100 nm. Palm vein recognition utilizes the characteristic that hemoglobin in blood absorbs near-infrared light, and a near-infrared-sensitive camera is used to obtain vein images under the irradiation of near-infrared light with a wavelength of 700 nm - 1000 nm. From the above information, it can be seen that the near-infrared rays adopted in the palm vein mode and the face mode are basically of the same wavelength, that is, the light sensitivity of the two modes is similar. It can be understood that if sharing an infrared light source affects the recognition effect of any one mode, it will also affect the other recognition mode.

[0111] For the intelligent lock provided in this application, at least two near-infrared light emitting diodes are distributed near each camera, that is, the face infrared light source and the palm vein infrared light source are respectively distributed near the visible light camera and the near-infrared camera on one side, and a shared infrared light source is also set between the visible light camera and the near-infrared camera. When a problem occurs with the near-infrared light emitting diodes distributed on one side of the visible light camera or the near-infrared camera, it only affects the shooting of the face module or the palm vein module. At this time, the palm vein module can continue to work normally. When a problem occurs with the shared near-infrared light emitting diodes set between the visible light camera and the near-infrared camera, neither the face mode nor the palm vein mode can work normally.

[0112] In this embodiment, the position of the light source with insufficient pixels, that is, the light source position information, can still be extracted from the image captured by the image edge information algorithm. It can be understood that when the light is insufficient, the edge information of the face or palm vein is also blurred and invisible, so as to confirm whether the light source type corresponding to the light source position information is a shared infrared light source or a single-sided light source. The single-sided light sources include the face infrared light source and the palm vein infrared light source.

[0113] Embodiment Five

[0114] For the intelligent lock control method based on image recognition provided in this embodiment, in step S1201, that is, to confirm the infrared light source position corresponding to the light source position information, it specifically includes the following steps:

[0115] S2011. If the first recognition module is a face module and the light source position information exists in the 2D image captured by visible light, then the light source type corresponding to the light source position information is the face infrared light source.

[0116] S2012. If the first recognition module is a face module and the light source position information exists in the depth image captured by the near-infrared camera, the light source type corresponding to the light source position information is a shared infrared light source.

[0117] S2013. If the first recognition module is a palm vein module and the light source position information is close to the shared infrared light source, the light source type corresponding to the light source position information is a shared infrared light source.

[0118] S2014. If the first recognition module is a palm vein module and the light source position information is close to the palm vein infrared light source, the light source type corresponding to the light source position information is a palm vein infrared light source.

[0119] Specifically, in this embodiment, the light-deficient position of the test image can be determined first. When the recognition module is a face module, the face image is synthesized by superimposing the 2D image captured by the visible light camera and the depth image captured by the near-infrared camera. When the 2D image is not clear but the depth image is clear, it can be determined that there is a problem with the face infrared light source. When the 2D image is clear but the depth image is not clear, it can be determined that there is a problem with the shared infrared light source.

[0120] When the recognition module is a palm vein module, if the light source position information in the palm vein image is close to the shared infrared light source, that is, the palm vein features near the palm vein infrared light source are almost blurred and invisible, while the features near the shared infrared light source are visible, it can be determined that there is a problem with the shared infrared light source. Similarly, if the light source position information in the palm vein image is close to the palm vein infrared light source, that is, the palm vein features near the palm vein infrared light source are almost blurred and invisible, while the features near the shared infrared light source are visible, it can be determined that there is a problem with the palm vein infrared light source.

[0121] Embodiment Six

[0122] The intelligent lock control method based on image recognition provided in this embodiment, after step S1201, that is, after confirming the infrared light source position corresponding to the light source position information, further specifically includes the following steps:

[0123] S6210. Perform a lens self-check on the camera corresponding to the infrared light source position during the set system self-check period. The lens self-check includes an occlusion self-check and a light self-check, and obtain the lens self-check result.

[0124] S6220. If the lens self-check result is a self-check failure, send a repair request message to the remote control terminal. The repair request message includes the occlusion position information and / or the light source position information.

[0125] Specifically, in this embodiment, the self-check time of the intelligent lock can be set, generally at night when users are all resting, such as 1 o'clock in the morning, etc. as the system self-check period. During this period, the intelligent lock can take pictures of the corresponding environmental information in front of the current camera by itself and perform information comparison, so as to realize occlusion self-check and light self-check, and timely obtain the device self-check information as the lens self-check result. After corresponding problems occur, self-cleaning or problem reporting, etc. can be set in time, rather than cleaning or problem reporting only when the user discovers the problem, improving the self-check intelligence and use reliability of this system.

[0126] Embodiment Seven

[0127] For the intelligent lock control method based on image recognition provided in this embodiment, before step S110, that is, before obtaining the first mode image through the first recognition module, as Figure 7 shown, it further specifically includes the following steps:

[0128] S1011. Detect the non-contact distance between the user and the intelligent lock through a distance sensor.

[0129] S1012. Determine at least one target camera required for the first recognition module according to the non-contact distance.

[0130] S1013. Collect the first mode image corresponding to the user through the target camera for feature recognition.

[0131] Specifically, when it is detected that the non-contact distance where the user is located is within the face recognition distance range, such as 0.5 meters to 2 meters, start the face module as the first recognition module, and start the corresponding visible light camera and near-infrared camera as the target cameras to collect the face features corresponding to the user; when the non-contact distance where the user is located is less than the distance within the face recognition distance range, start the palm vein module as the first recognition module, and start the corresponding near-infrared camera as the target camera to collect the palm vein features corresponding to the user.

[0132] For the intelligent lock control method based on image recognition provided in this embodiment, by timely analyzing the reasons for the biometric feature images that do not meet the recognition requirements obtained during shooting, it can be determined that when the camera is occluded, such as being dirty, it can be timely switched to other effective biometric recognition modes, enhancing the anti-interference ability of this system while improving the recognition effectiveness and timeliness of biometric features, and maintaining the recognition stability of this system.

[0133] It should be understood that the magnitudes of the sequence numbers of the steps in the above embodiments do not mean the order of execution. The execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0134] Embodiment Eight

[0135] In one embodiment, an intelligent lock control device based on image recognition is provided. The intelligent lock control device based on image recognition corresponds one-to-one with the intelligent lock control method based on image recognition in the above embodiment. As Figure 8 shown, the intelligent lock control device based on image recognition includes a first preprocessed image acquisition module 110, a cause analysis result acquisition module 120, a camera type confirmation module 130, a command sending module 140, and a feature recognition result acquisition module 150. The detailed description of each functional module is as follows:

[0136] The first preprocessed image acquisition module 110 is configured to obtain a first pattern image through a first recognition module, perform image preprocessing on the first pattern image, and obtain a first preprocessed image. Among them, the first recognition module is any one of a face module or a palm vein module, and the second recognition module is the other module different from the first recognition module among the face module or the palm vein module.

[0137] The cause analysis result acquisition module 120 is configured to, if the first preprocessed image is an image with unclear features, perform unclear cause analysis on the image with unclear features to obtain a cause analysis result.

[0138] The camera type confirmation module 130 is configured to, if the cause analysis result is that the lens is blocked, record the occlusion position information corresponding to the occlusion, confirm the camera type corresponding to the occlusion position information, and thus confirm whether it affects the second recognition module.

[0139] The command sending module 140 is configured to, if the first recognition module is a face module and the camera type is a near-infrared camera, send a switching recognition module command to receive the third pattern features input by the user through any one of the other recognition modules as the third recognition module.

[0140] The feature recognition result acquisition module 150 is configured to perform feature recognition on the third pattern features, obtain a feature recognition result, and confirm the opening of the intelligent lock according to the feature recognition result.

[0141] Furthermore, the intelligent lock control device specifically further includes the following modules:

[0142] The switching recognition module command sending module 1301 is configured to, if the first recognition module is a face module and the camera type is a visible light camera, send a switching recognition module command to receive the third pattern features input by the user through any one of the other recognition modules as the third recognition module, or receive the second pattern features input by the user through the second recognition module.

[0143] The switching recognition module instruction sending module 1302 is configured to send a switching recognition module instruction if the first recognition module is a palm vein module and the camera type is a near-infrared camera, so as to receive third mode features input by the user through any one of other recognition modules as the third recognition module.

[0144] The intelligent lock opening confirmation module 1303 is configured to perform feature recognition on the third mode features or the second mode features, obtain a feature recognition result, and confirm the opening of the intelligent lock according to the feature recognition result.

[0145] Furthermore, the intelligent lock control device further specifically includes the following modules:

[0146] The cleaning instruction sending module 1401 is configured to send a panel cleaning instruction to clean the transparent panel corresponding to the near-infrared camera.

[0147] The information change result determining module 1402 is configured to determine the information change result of the occlusion position information after a preset cleaning waiting time.

[0148] The step repeating execution module 1403 is configured to repeat the step of obtaining the first mode image by the first recognition module if the information change result is a decrease in the position information.

[0149] The repair information sending module 1404 is configured to send repair information to the remote control end if the information change result is that the position information has no change, and the repair information includes the occlusion position information.

[0150] Furthermore, the intelligent lock control device further specifically includes the following modules:

[0151] The intelligent lock opening confirmation module 1101 is configured to perform feature recognition based on the feature clear image if the first preprocessed image is a feature clear image, obtain a feature recognition result, and confirm the opening of the intelligent lock according to the feature recognition result.

[0152] Furthermore, the intelligent lock control device further specifically includes the following modules:

[0153] The infrared light source position confirmation module 1201 is configured to record the light source position information corresponding to the light source with insufficient light if the cause analysis result is insufficient light of the lens, confirm the infrared light source position corresponding to the light source position information, so as to confirm whether it affects the second recognition module.

[0154] The switching recognition module instruction sending module 1202 is configured to send a switching recognition module instruction if the first recognition module is a face module and the infrared light source position is a face infrared light source, so as to receive third mode features input by the user through any one of other recognition modules as the third recognition module, or receive second mode features input by the user through the second recognition module.

[0155] The switching recognition module instruction sending module 1203 is configured to send a switching recognition module instruction if the first recognition module is a palm vein module and the infrared light source position is a shared infrared light source, so as to receive third-mode features input by the user through any one of the other recognition modules as the third recognition module.

[0156] The intelligent lock opening confirmation module 1204 is configured to perform feature recognition on the third-mode features or the second-mode features, obtain a feature recognition result, and confirm the opening of the intelligent lock according to the feature recognition result.

[0157] Furthermore, the intelligent lock control device specifically further includes the following modules:

[0158] The lens self-checking module 6210 is configured to perform a lens self-check on the camera corresponding to the infrared light source position during a set system self-check period. The lens self-check includes an occlusion self-check and a light self-check, and obtain a lens self-check result.

[0159] The warranty information sending module 6220 is configured to send a repair request information to a remote control end if the lens self-check result is a self-check failure. The repair request information includes occlusion position information and / or light source position information.

[0160] Furthermore, the intelligent lock control device specifically further includes the following modules:

[0161] The non-contact distance detection module 1011 is configured to detect the non-contact distance between the user and the intelligent lock through a distance sensor.

[0162] The target camera determination module 1012 is configured to determine at least one target camera required to be used by the first recognition module according to the non-contact distance.

[0163] The feature recognition module 1013 is configured to collect a first-mode image corresponding to the user through the target camera for feature recognition.

[0164] For the specific limitations on the intelligent lock control device based on image recognition, reference can be made to the limitations on the intelligent lock control method based on image recognition in the above text, which will not be elaborated here. Each module in the above intelligent lock control device based on image recognition can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in or independent of the processor in the electronic device in the form of hardware, or stored in the memory in the electronic device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0165] Embodiment Nine

[0166] This embodiment further provides an electronic device, such as Figure 9As shown in the figure, it includes: a microprocessor that supports at least two biometric identifications, and a functional module electrically connected to the microprocessor. The functional module includes: a camera module, an infrared illumination light source, a face module, a palm vein module, a distance sensing module, a light sensing module, a memory module, an Ethernet communication module, a speaker, an electromagnetic lock control module, and a power supply module. The Ethernet communication module is connected to a remote control terminal through Ethernet.

[0167] The distance sensing module is used to detect whether there is an object within a preset recognition distance. The electromagnetic lock control module is used to perform an unlocking operation.

[0168] The light sensing module is used to detect environmental light information and send the environmental light information to the microprocessor, so that the microprocessor switches the working mode of the camera in the camera module according to the environmental light information.

[0169] The camera module is arranged on the lower side of the transparent panel and includes at least one visible light camera and at least one near-infrared camera. A visible light filter is embedded in the transparent panel. At least two near-infrared light emitting tubes are evenly distributed around each camera, and at least one near-infrared light emitting tube is evenly distributed between two cameras. The near-infrared light emitting tubes are used to provide infrared fill light function for the cameras in the camera module.

[0170] The microprocessor controls the camera module and the near-infrared light emitting tubes to identify a face or a palm vein and to identify the face mode image or palm vein mode image captured by the camera module.

[0171] When the microprocessor executes a computer program, it implements the intelligent lock control method based on image recognition in the above embodiment, such as Figure 2 Steps S110 to S150 shown in the figure. Alternatively, when the microprocessor executes a computer program, it implements the functions of each module / unit of the intelligent lock control device based on image recognition in the above embodiment, such as Figure 9 The functions of module 110 to module 150 shown in the figure. To avoid repetition, it will not be elaborated here.

[0172] Embodiment Ten

[0173] This embodiment also provides an electronic device, which specifically further includes: a lens self-cleaning mechanism arranged on the transparent panel for cleaning the camera area.

[0174] Specifically, the intelligent lock provided in this embodiment has a transparent panel with an automatic cleaning function. This panel is integrated with a miniature lens self-cleaning device, which can automatically clean the transparent panel before recognition to remove water droplets and dirt. The lens self-cleaning device, in cooperation with the intelligent lock control system based on image recognition provided in this embodiment, can intelligently adjust the frequency and intensity of automatic cleaning according to environmental conditions and usage frequency; and after receiving the panel cleaning instruction, it can start the lens self-cleaning device to clean the corresponding position on the transparent panel of the camera that has recently become dirty. To speed up the unlocking speed, at this time, a full-panel cleaning can be performed as needed, or only the corresponding position on the transparent panel of the camera that has recently become dirty can be cleaned.

[0175] Those of ordinary skill in the art can understand that all or part of the processes of implementing the methods in the above embodiments can be completed by instructing relevant hardware through a computer program. This computer program can be stored in a non-volatile computer-readable medium. When this computer program is executed, it can include the processes of the embodiments of the above various methods. Among them, any reference to a memory, storage, database, or other medium used in the various embodiments of this application can include non-volatile and / or volatile memories. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or an external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0176] Those skilled in the art can clearly understand that for the convenience and brevity of description, only the above-mentioned division of each functional unit and module is used as an example. In actual applications, the above functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0177] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. An intelligent lock control method based on image recognition, characterized in that, Including: Obtain a first mode image through a first recognition module, perform image preprocessing on the first mode image to obtain a first preprocessed image, where the first recognition module is any one of a face module or a palm vein module, and the second recognition module is the other module different from the first recognition module in the face module or the palm vein module; If the first preprocessed image is an image with unclear features, perform an analysis of the unclear reason on the image with unclear features to obtain a reason analysis result; If the reason analysis result is that the lens is blocked, record the occlusion position information corresponding to the occlusion, confirm the camera type corresponding to the occlusion position information, so as to confirm whether it affects the second recognition module; If the first recognition module is a face module and the camera type is a near-infrared camera, send a switching recognition module instruction to receive a third mode feature input by the user through any one of other recognition modules as a third recognition module; Perform feature recognition on the third mode feature to obtain a feature recognition result, and confirm the opening of the intelligent lock according to the feature recognition result; If the first recognition module is a face module and the camera type is a visible light camera, send a switching recognition module instruction to receive a third mode feature input by the user through any one of other recognition modules as a third recognition module, or receive a second mode feature input by the user through the second recognition module; if the first recognition module is a palm vein module and the camera type is a near-infrared camera, send a switching recognition module instruction to receive a third mode feature input by the user through any one of other recognition modules as a third recognition module; perform feature recognition on the third mode feature or the second mode feature to obtain a feature recognition result, and confirm the opening of the intelligent lock according to the feature recognition result; Send a panel cleaning instruction to clean the transparent panel corresponding to the near-infrared camera; after a preset cleaning waiting time, determine the information change result of the occlusion position information; if the information change result is a decrease in the position information, repeat the step of obtaining the first mode image by the first recognition module; if the information change result is that the position information does not change, send a repair information to the remote control terminal, and the repair information includes the occlusion position information.

2. The intelligent lock control method based on image recognition according to claim 1, characterized in that After obtaining the reason analysis result, it further includes: If the reason analysis result is that the lens has insufficient light, record the light source position information corresponding to the light source with insufficient light, confirm the infrared light source position corresponding to the light source position information, so as to confirm whether it affects the second recognition module; If the first recognition module is a face module and the infrared light source position is a face infrared light source, send a switching recognition module instruction to receive a third mode feature input by the user through any one of other recognition modules as a third recognition module, or receive a second mode feature input by the user through the second recognition module; If the first recognition module is a face recognition module and the position of the infrared light source is a shared infrared light source, send a recognition module switching instruction to receive the third mode features input by the user through any one of the other recognition modules as the third recognition module; If the first recognition module is a palm vein recognition module and the position of the infrared light source is a shared infrared light source, send a recognition module switching instruction to receive the third mode features input by the user through any one of the other recognition modules as the third recognition module; If the first recognition module is a face recognition module and the position of the infrared light source is a palm vein infrared light source, send a recognition module switching instruction to receive the third mode features input by the user through any one of the other recognition modules as the third recognition module, or receive the second mode features input by the user through the second recognition module; Perform feature recognition on the third mode features or the second mode features to obtain a feature recognition result, and confirm the opening of the smart lock according to the feature recognition result.

3. The intelligent lock control method based on image recognition according to claim 2, wherein The confirmation of the infrared light source position corresponding to the light source position information includes: If the first recognition module is a face recognition module and the light source position information exists on the 2D image captured by visible light, the light source type corresponding to the light source position information is a face infrared light source; If the first recognition module is a face recognition module and the light source position information exists on the depth image captured by the near-infrared camera, the light source type corresponding to the light source position information is a shared infrared light source; If the first recognition module is a palm vein recognition module and the light source position information is close to the shared infrared light source, the light source type corresponding to the light source position information is a shared infrared light source; If the first recognition module is a palm vein recognition module and the light source position information is close to the palm vein infrared light source, the light source type corresponding to the light source position information is a palm vein infrared light source.

4. The intelligent lock control method based on image recognition according to claim 2, characterized in that, After the confirmation of the infrared light source position corresponding to the light source position information, it further includes: Perform a lens self-check on the camera corresponding to the infrared light source position during the set system self-check period. The lens self-check includes occlusion self-check and light self-check to obtain a lens self-check result; If the lens self-check result is a self-check failure, send a repair request message to the remote control terminal. The repair request message includes the occlusion position information and / or the light source position information.

5. The intelligent lock control method based on image recognition according to claim 1, wherein Before obtaining the first mode image through the first recognition module, it further includes: Detect the non-contact distance between the user and the smart lock through a distance sensor; Determine at least one target camera required for the first recognition module according to the non-contact distance; Collect the first mode image corresponding to the user through the target camera for feature recognition.

6. An intelligent lock control device based on image recognition, characterized in that, It includes: A first preprocessing image acquisition module, which is used to obtain a first mode image through a first recognition module, perform image preprocessing on the first mode image, and obtain a first preprocessing image. Among them, the first recognition module is any one of a face recognition module or a palm vein recognition module, and the second recognition module is the other module different from the first recognition module in the face recognition module or the palm vein recognition module; A cause analysis result acquisition module, configured to perform unexplained cause analysis on the image with unclear features if the first preprocessed image is an image with unclear features, and acquire a cause analysis result; A camera type confirmation module, configured to record the occlusion position information corresponding to the occlusion if the cause analysis result is that the lens is occluded, confirm the camera type corresponding to the occlusion position information, so as to confirm whether it affects the second recognition module; An instruction sending module, configured to send a recognition module switching instruction if the first recognition module is a face module and the camera type is a near-infrared camera, so as to receive third-mode features input by the user through any one of other recognition modules as a third recognition module; A feature recognition result acquisition module, configured to perform feature recognition on the third-mode features, acquire a feature recognition result, and confirm the opening of the intelligent lock according to the feature recognition result; if the first recognition module is a face module and the camera type is a visible light camera, send a recognition module switching instruction, so as to receive third-mode features input by the user through any one of other recognition modules as a third recognition module, or receive second-mode features input by the user through the second recognition module; if the first recognition module is a palm vein module and the camera type is a near-infrared camera, send a recognition module switching instruction, so as to receive third-mode features input by the user through any one of other recognition modules as a third recognition module; perform feature recognition on the third-mode features or the second-mode features, acquire a feature recognition result, and confirm the opening of the intelligent lock according to the feature recognition result; send a panel cleaning instruction to clean the transparent panel corresponding to the near-infrared camera; after a preset cleaning waiting time, determine the information change result of the occlusion position information; if the information change result is a reduction in the position information, repeat the step of the first recognition module acquiring the first-mode image; if the information change result is that the position information does not change, send a repair request information to a remote control terminal, where the repair request information includes the occlusion position information.

7. An electronic device, characterized in that, Including: A microprocessor supporting at least two biometric identifications, and a function module electrically connected to the microprocessor, where the function module includes: a camera module, an infrared illumination light source, a face module, a palm vein module, a distance sensing module, a light sensing module, a memory module, an Ethernet communication module, a speaker, an electromagnetic lock control module, and a power supply module, and the Ethernet communication module is connected to the remote control terminal through Ethernet; A memory module, configured to store preset feature data for at least two biometric identifications and a computer program that can run on the microprocessor, and the computer program is executed by the microprocessor, so that the microprocessor implements the intelligent lock control method based on image recognition according to any one of claims 1-5; the microprocessor controls the camera module and the near-infrared light emitting tube to perform recognition on a face or a palm vein; The light-sensitive induction module is used to detect ambient light information and send the ambient light information to the microprocessor, so that the microprocessor switches the working mode of the camera in the camera module according to the ambient light information; The camera module is arranged under the transparent panel and includes at least one visible light camera and at least one near-infrared camera. A visible light filter is embedded in the transparent panel; at least two of the near-infrared light emitting tubes are evenly distributed around each camera, and at least one of the near-infrared light emitting tubes is evenly distributed between the two cameras. The near-infrared light emitting tubes are used to provide infrared fill light function for the cameras in the camera module.

8. The electronic device according to claim 7, wherein It further includes: A lens self-cleaning mechanism arranged on the transparent panel for cleaning the camera area.

Images