Frame lamp illumination control method, device and system applied to safety helmet

Abstract

The invention is suitable for the technical field of personal safety protection equipment, and provides a frame lamp illumination control method, device and system applied to a safety helmet, and the method comprises the steps: obtaining head movement data representing the head posture of a wearer in real time, and synchronously obtaining eye movement data representing the eyeball sight of the wearer; carrying out data fusion calculation on the head movement data and the eye movement data so as to solve an absolute sight line vector which points to a visual focus of the wearer under a preset world coordinate system and has compensated head movement interference; performing time domain and space domain conjoint analysis on a time sequence data stream formed by the continuous absolute sight line vectors so as to identify the current eye movement behavior of the wearer as one of at least two preset eye movement modes; and according to the identified eye movement mode, executing a differential light source driving control strategy corresponding to the identified eye movement mode. According to the invention, the problem that accurate illumination cannot be carried out by accurately judging the intention of a user in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of personal safety protection equipment, and in particular to a lighting control method, device and system for a bracket lamp applied to a safety helmet. Background Art

[0002] In environments requiring helmets, such as construction, mining, tunneling, emergency rescue, nighttime inspections, and various types of industrial maintenance, workers often work in low-light conditions. To ensure worker safety and improve work efficiency, helmets now include a built-in headlamp or mining lamp. The key advantage of this head-mounted lighting device is that it frees the user's hands and directs the beam roughly in front of their field of vision.

[0003] However, existing helmet lighting solutions generally have one or more of the following significant technical defects and application pain points:

[0004] The beam direction of most traditional bracket lights is fixed and rigidly bound to the orientation of the helmet, which means that the light can only illuminate the direction directly facing the wearer's head. However, in actual work, the visual focus of the human eye (the direction the eyes are looking) is often not completely consistent with the orientation of the head, and it may frequently shift independently of head movement. When the user needs to observe an object not directly in front of them (such as looking down at a drawing or looking at a tool sideways), but the headlamp is still facing directly forward, the light at the focal point of the line of sight is insufficient, affecting work efficiency and accuracy. Therefore, the user must turn their head to guide the beam, resulting in the light not being able to cover the actual visual focus in real time and accurately. This not only reduces comfort, but may also cause the user to ignore the surrounding environment due to frequent turning of the head, posing a safety hazard.

[0005] To address these issues, some headlamps with manually adjustable angles have appeared on the market. These products typically require the user to manually adjust the lamp head to change its pitch or yaw angle. While this provides a certain degree of flexibility, its drawbacks are also obvious in many critical work scenarios. When workers need to hold tools with both hands, climb ladders, or operate in confined spaces, freeing one hand to adjust the light becomes extremely inconvenient or even impossible, directly leading to reduced work efficiency. Furthermore, frequent manual adjustments can distract the user and pose a safety hazard. Summary of the Invention

[0006] Based on this, the purpose of the present invention is to provide a method, device and system for controlling the lighting of a bracket lamp applied to a helmet, so as to fundamentally solve the problem that the existing system cannot accurately judge the user's intention and provide precise lighting.

[0007] According to an embodiment of the present invention, a method for controlling a lighting fixture of a safety helmet includes:

[0008] Acquire head movement data representing the wearer's head posture in real time, and simultaneously acquire eye movement data representing the wearer's eye gaze;

[0009] Performing data fusion calculation on the head movement data and the eye movement data to calculate an absolute sight line vector pointing to the wearer's visual focus in a preset world coordinate system and compensating for head movement interference;

[0010] Performing a joint time-domain and space-domain analysis on a time-series data stream consisting of continuous absolute gaze vectors to identify the wearer's current eye movement behavior as one of at least two preset eye movement patterns, wherein the eye movement pattern includes at least a fixation pattern representing a stable visual focus and a saccade pattern representing a rapid visual sweep;

[0011] According to the identified eye movement pattern, a differentiated light source drive control strategy corresponding to the identified eye movement pattern is executed, wherein the differentiated light source drive control strategy includes:

[0012] When the recognized eye movement pattern is a gaze pattern, a target lighting point is calculated based on a set of absolute sight line vectors constituting the recognized gaze pattern, and the rotatable light source on the stand lamp is driven to adjust its projection angle relative to the helmet in real time in a manner to offset the wearer's head movement, so as to project the light beam of the light source toward the target lighting point;

[0013] When the identified eye movement pattern is a scanning pattern, a light source angle suppression strategy is executed, wherein the light source angle suppression keeps the projection angle of the rotatable light source relative to the helmet stable during the scanning period.

[0014] In addition, the lighting control method for a safety helmet bracket lamp according to the above embodiment of the present invention may also have the following additional technical features:

[0015] Furthermore, the step of synchronously acquiring eye movement data representing the wearer's eye sight includes:

[0016] Using a near-infrared light source to project infrared light toward the wearer's eye area, so as to form at least two relatively fixed Purchinye light spots on the corneal surface;

[0017] Continuously capturing a sequence of eye images including the pupil outline and the Purchinye light spot using an image sensor;

[0018] In each frame of the eye image sequence, the geometric center of the pupil and the geometric centers of at least two Purchinye spots are located by an image processing algorithm;

[0019] A differential vector pointing from the geometric center of each Purchinye light spot to the geometric center of the pupil is calculated, and the calculated differential vectors are fused to generate the eye movement data.

[0020] Furthermore, the step of locating the geometric center of the pupil and the geometric centers of at least two Purchinye light spots by using an image processing algorithm includes:

[0021] In each frame of the image, the geometric center of all Purchinye spots is identified and located by a fast threshold segmentation algorithm;

[0022] Based on the geometric centers of the multiple located Purchinye light spots, a region of interest containing only the pupil is delineated in each frame image;

[0023] In the region of interest, the pupil contour is identified according to a contour detection algorithm, and the pupil contour is fitted using an ellipse fitting algorithm to determine the geometric center of the pupil.

[0024] Furthermore, the step of performing data fusion calculation on the head movement data and the eye movement data to calculate an absolute sight line vector pointing to the wearer's visual focus in a preset world coordinate system and compensating for head movement interference includes:

[0025] Processing the head motion data through a posture solving algorithm to derive a rotation matrix or quaternion describing the real-time rotation of the helmet body coordinate system relative to the preset world coordinate system;

[0026] Converting the eye movement data into a sight unit direction vector in the helmet body coordinate system through a pre-calibrated mapping model;

[0027] The rotation matrix or quaternion is applied to the sight unit direction vector in the helmet body coordinate system, and the absolute sight vector in the preset world coordinate system is finally obtained through coordinate system transformation operation.

[0028] Furthermore, the step of performing a joint time domain and space domain analysis on the time series data stream consisting of the continuous absolute gaze vectors to identify the wearer's current eye movement behavior as one of at least two preset eye movement patterns includes:

[0029] Calculating the angular velocity of the absolute sight line vector in real time based on the time series data stream;

[0030] comparing the calculated angular velocity with at least two preset velocity thresholds, the velocity thresholds comprising a higher glance velocity threshold and a lower gaze velocity threshold;

[0031] When the angular velocity is higher than the saccadic velocity threshold, the current eye movement behavior is identified as a saccadic mode;

[0032] When the angular velocity is lower than the gaze velocity threshold, and its duration exceeds a preset gaze time threshold, and the spatial discreteness of the absolute gaze vector within its duration is lower than a preset spatial dispersion threshold, the current eye movement behavior is identified as a gaze pattern.

[0033] Furthermore, the step of calculating a target lighting point based on a set of absolute sight line vectors constituting the gaze pattern identified this time includes:

[0034] Extracting candidate gaze clusters that are identified as gaze patterns and consist of a set of absolute gaze vectors;

[0035] The arithmetic mean or weighted mean of the coordinates of all vector endpoints in the candidate fixation cluster is calculated to determine the coordinates representing the geometric center of the candidate fixation cluster, and the coordinates of the geometric center of the candidate fixation cluster are determined as the target illumination point.

[0036] Furthermore, the step of driving the rotatable light source on the bracket lamp to adjust its projection angle relative to the helmet in real time in a manner to offset the wearer's head movement so as to project the light source beam toward the target lighting point includes:

[0037] The target lighting point in the preset world coordinate system is converted into the target projection vector in the helmet body coordinate system through an inverse coordinate system transformation based on real-time head motion data;

[0038] Decomposing the target projection vector into a target pitch angle and a target yaw angle required by the rotatable light source through an inverse kinematics model;

[0039] comparing the target pitch angle and target yaw angle with the current actual pitch angle and yaw angle in the head motion data to calculate an angle error;

[0040] A PID controller is used to take the angle error as input and continuously generate a control signal for driving a motor of the rotatable light source until the angle error converges to a preset range.

[0041] Furthermore, the eye movement mode also includes a smooth tracking mode representing smooth visual pursuit of a dynamic target, and the differentiated light source drive control strategy also includes:

[0042] When the identified eye movement pattern is a smooth tracking pattern, a dynamically moving target path is calculated based on the absolute line of sight vector sequence that constitutes the identified smooth tracking pattern, and the rotatable light source is driven to adjust its projection angle relative to the helmet in real time to smoothly move the light source beam along the target path at a matching speed.

[0043] Another embodiment of the present invention is to provide a lighting control device for a safety helmet bracket lamp, the device comprising:

[0044] A multi-source data acquisition module is used to acquire head movement data representing the wearer's head posture in real time, and simultaneously acquire eye movement data representing the wearer's eye sight;

[0045] an absolute sight line vector calculation module, configured to perform data fusion calculation on the head movement data and the eye movement data to calculate an absolute sight line vector pointing to the wearer's visual focus in a preset world coordinate system and compensating for head movement interference;

[0046] an eye movement pattern recognition module, configured to perform a joint temporal and spatial analysis on a time-series data stream consisting of continuous absolute gaze vectors to identify the wearer's current eye movement behavior as one of at least two preset eye movement patterns, wherein the eye movement pattern includes at least a gaze pattern representing a stable visual focus and a saccade pattern representing a rapid visual sweep;

[0047] The lighting control module is configured to execute a differentiated light source drive control strategy corresponding to the identified eye movement pattern according to the identified eye movement pattern, wherein the differentiated light source drive control strategy includes:

[0048] When the recognized eye movement pattern is a gaze pattern, a target lighting point is calculated based on a set of absolute sight line vectors constituting the recognized gaze pattern, and the rotatable light source on the stand lamp is driven to adjust its projection angle relative to the helmet in real time in a manner to offset the wearer's head movement, so as to project the light beam of the light source toward the target lighting point;

[0049] When the identified eye movement pattern is a scanning pattern, a light source angle suppression strategy is executed, wherein the light source angle suppression keeps the projection angle of the rotatable light source relative to the helmet stable during the scanning period.

[0050] Another embodiment of the present invention is to provide a lighting control system for a safety helmet bracket lamp, the system comprising:

[0051] helmet;

[0052] A head posture sensor, mounted on the helmet, for acquiring head movement data of the wearer;

[0053] an eye tracking module, integrated into the helmet or its mounting accessories, for acquiring eye movement data of the wearer;

[0054] a bracket lamp, mounted on the safety helmet, comprising a bracket and a rotatable light source;

[0055] A central processing unit is electrically connected to the head posture sensor, the eye tracking module and the bracket light, and the central processing unit is configured to execute the bracket light lighting control method applied to the helmet as described above.

[0056] The embodiment of the present invention provides a lighting control method for a bracket lamp applied to a helmet. By synchronously acquiring head movement data and eye movement data in real time and performing data fusion calculation on the two, the absolute sight line vector that compensates for the interference of head movement is solved, thereby realizing that the lighting control system accurately locks the wearer's real visual focus, and solves the problem that the traditional headlamp light beam can only follow the physical orientation of the head but cannot be aligned with the wearer's real sight line. It ensures that no matter how the wearer's head posture changes, the light beam can always be accurately projected to the target point he is observing; by jointly analyzing the time domain and spatial domain of the time series data stream composed of the absolute sight line vector to identify the specific eye movement pattern and executing the corresponding differentiated light source drive control strategy, the wearer's subjective intention is intelligently judged, and the frequent and ineffective shaking of the light beam caused by unconscious and rapid scanning of the line of sight is effectively avoided, thereby solving the problem of the existing automatic headlights being unable to focus on the wearer. The system cannot distinguish user intentions, which can easily cause visual interference and wearing discomfort. However, when a gaze pattern is identified, the system calculates the precise target lighting point based on a set of absolute line of sight vectors that constitute the gaze, and drives the rotatable light source for directional projection, ensuring that the lighting energy is efficiently concentrated on the key areas that the wearer is currently most interested in. When a scanning pattern is identified, the system implements a light source angle suppression strategy to keep the light source projection angle stable relative to the helmet during the scanning period, greatly enhancing the wearer's visual comfort and safety when observing a large environment or searching for a target. By driving the rotatable light source to adjust the projection angle to offset the wearer's head movement, the light beam is absolutely stable in the world coordinate system. When the wearer looks at a fixed point and turns their head at the same time, the light beam can always lock onto the sign without shaking with the head, ensuring continuous, uninterrupted and clear illumination of the target. This solves the problem of the existing inability to accurately identify user intentions and provide precise lighting. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 1 is a flow chart of a method for controlling a lighting system of a safety helmet support lamp according to a first embodiment of the present invention;

[0058] Figure 2 for Figure 1Specific flow diagram of step S10;

[0059] Figure 3 Schematic diagram of the structure of a lighting control device for a support lamp applied to a helmet in a second embodiment of the present invention;

[0060] Figure 4 Schematic diagram of the structure of a bracket lamp lighting control system applied to a helmet in a third embodiment of the present invention;

[0061] The following specific embodiments will further illustrate the present invention in conjunction with the above-mentioned drawings. DETAILED DESCRIPTION

[0062] To facilitate understanding of the present invention, the present invention will be described more fully below with reference to the accompanying drawings. The drawings illustrate several embodiments of the present invention. However, the present invention may be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and comprehensive understanding of the present invention.

[0063] It should be noted that when an element is referred to as being "fixed to" another element, it may be directly on the other element or there may be an intermediate element. When an element is referred to as being "connected to" another element, it may be directly connected to the other element or there may be an intermediate element. The terms "vertical," "horizontal," "left," "right," and similar expressions used herein are for illustrative purposes only.

[0064] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0065] Example 1

[0066] See also Figure 1 , which shows a lighting control method for a safety helmet bracket lamp in a first embodiment of the present invention. For ease of description, only the parts related to the embodiment of the present invention are shown. The lighting control method for a safety helmet bracket lamp provided by the embodiment of the present invention includes:

[0067] Step S10, acquiring head movement data representing the wearer's head posture in real time, and synchronously acquiring eye movement data representing the wearer's eye sight;

[0068] Among them, in one embodiment of the present invention, the method is applied to the stand light lighting control system of the safety helmet, wherein the stand light lighting control system includes a safety helmet, a head posture sensor, an eye tracking module, a stand light, and a central processing unit, wherein the embodiment of the present invention is mainly used to control the central processing unit to implement the stand light lighting control method described in this embodiment, wherein when the wearer puts on the safety helmet and turns on the power, the central processing unit sends a request to the head posture sensor (IMU) at a fixed frequency (such as 100Hz or 200Hz) through the I2C bus, and reads the original three-axis acceleration and three-axis angular velocity data output by it in real time as head motion data. At the same time, its central processing unit will obtain a high-precision timestamp from the hardware timer and package the head motion data together with the timestamp. While obtaining the head motion data, the central processing unit controls the eye tracking module through the SPI interface to obtain eye data. Specifically, in one embodiment of the present invention, refer to Figure 2 As shown, the steps of synchronously acquiring eye movement data representing the wearer's eye sight include:

[0069] Step S11, using a near-infrared light source to project infrared light toward the wearer's eye area, so as to form at least two relatively fixed Purchinye light spots on the corneal surface;

[0070] Step S12, using an image sensor to continuously capture a sequence of eye images including pupil contours and Purchinye spots;

[0071] Step S13, in each frame of the eye image sequence, locating the geometric center of the pupil and the geometric centers of at least two Purchinye spots using an image processing algorithm;

[0072] Step S14 , calculating a differential vector from the geometric center of each Purchinye light spot to the geometric center of the pupil, and fusing the calculated differential vectors to generate eye movement data.

[0073] Specifically, the eye tracking module includes an infrared LED light source and a near-infrared (NIR) camera. The infrared LED light source uses two 850nm high-intensity infrared LEDs, which are asymmetrically mounted around the NIR camera lens. The specific layout of the infrared LED light sources can be, for example, one located 10mm directly above the camera and the other 8mm below and to the left of the camera. This asymmetrical layout ensures that the displacement vectors of the two Purchinye spots on the infrared camera's image sensor are linearly independent when the eye moves. The NIR camera uses a CMOS image sensor with a global shutter, which prevents image smear and the rolling shutter effect caused by rapid eye movements (saccades). The NIR camera module is equipped with a lens with an infrared passband filter to filter out visible light interference, allowing only infrared light near 850nm to pass. The central processing unit first controls the two infrared LEDs in the eye tracking module to continuously illuminate, projecting invisible near-infrared light toward the wearer's eye area, forming two stable Purchinye spots on the cornea. At the same time, its near-infrared camera continuously captures an eye image sequence containing the pupil and the light spot at a preset frame rate (for example, 120fps). At this time, the central processing unit continuously receives the eye image sequence data from the near-infrared camera, and then performs image preprocessing on the eye image sequence data, including noise reduction processing and contrast enhancement processing, wherein the noise reduction processing is to apply a Gaussian blur kernel to the original image for slight smoothing processing to remove the random noise generated by the sensor; the contrast enhancement processing is to apply a local histogram equalization algorithm to enhance the contrast between the pupil area and the iris and sclera, especially under poor lighting conditions. Furthermore, an image processing algorithm is performed on each frame of the eye image sequence to locate the geometric center of the pupil and the geometric center of at least two Purchinje light spots. Then, a differential vector pointing from the geometric center of each Purchinye light spot to the geometric center of the pupil is calculated, where the direction and length of these two differential vectors are directly related to the rotation angle of the eyeball. Furthermore, the calculated differential vectors are fused to obtain eye movement data. Before fusion, the central processing unit evaluates the signal quality of each light spot, where the evaluation indicators include the brightness, roundness and area of ​​the light spot. At this time, the central processing unit calculates a quality score between 0 and 1 for each light spot. For example, if the light spot corresponding to an infrared LED light source is partially blocked by the eyelid, its area and roundness will decrease, resulting in a decrease in its quality score. The final eye movement data is obtained by weighted averaging the two differential vectors. At this time, if the signal quality of a light spot is poor, its contribution to the final result will be automatically reduced, so that the output eye movement data is extremely robust to instantaneous interference such as blinking and partial occlusion. At this time, the central processing unit uses the calculated two-dimensional vector as the eye movement data of this frame.At the same time, it obtains a high-precision timestamp from the hardware timer and packages the eye movement data together with the timestamp. At this time, the central processing unit synchronizes each set of head movement data and eye movement data by adding a precise hardware timestamp.

[0074] Furthermore, in one embodiment of the present invention, the step of locating the geometric center of the pupil and the geometric centers of at least two Purchinye spots using an image processing algorithm includes:

[0075] In each frame of the image, the geometric center of all Purchinye spots is identified and located by a fast threshold segmentation algorithm;

[0076] Based on the geometric centers of the multiple located Purchinye light spots, a region of interest containing only the pupil is delineated in each frame image;

[0077] In the region of interest, the pupil contour is identified according to the contour detection algorithm, and the pupil contour is fitted by the ellipse fitting algorithm to determine the geometric center of the pupil.

[0078] Specifically, since the Purchinye spot is the brightest area in the image, the central processing unit uses a fast threshold segmentation algorithm to identify and locate the geometric center of all Purchinye spots. First, the central processing unit sets a relatively high fixed brightness threshold (for example, a grayscale value of 230) to binarize the image. At this time, all pixels above this threshold are set to 1, and the rest are set to 0. Then a connected component analysis is performed on the binary image to find all independent white pixel clusters. The central processing unit then filters out valid Purchinye spots based on a preset area range (for example, 5 to 50 pixels) and roundness (>0.7). At this time, for each valid spot that passes the screening, the central processing unit traverses all the pixels it contains and calculates its respective pixel centroid (i.e., the geometric center) to obtain the precise center coordinates of the geometric center.

[0079] Furthermore, the central processing unit dynamically calculates a rectangular region of interest (ROI) based on the central coordinates of the geometric centers of the multiple located Purchinye light spots and the pre-calibrated typical geometric relationship between the pupil and the light spot. This rectangular ROI can frame the pupil with a high probability, thus avoiding searching for the pupil across the entire image and significantly reducing the computational effort. A contour detection algorithm (such as the Canny edge detection algorithm) is then applied only within the rectangular ROI. The threshold of this algorithm is adaptive and dynamically adjusted based on the grayscale histogram within the rectangular ROI to accurately outline the pupil's edge contour. After detection using the contour detection algorithm, a set of edge points representing the pupil contour is obtained. Because the human eye projects a circular pupil as an ellipse at an oblique viewing angle, the pupil typically appears elliptical in the image. Therefore, ellipse fitting is the most robust method for determining its center. The central processing unit then uses the edge point set detected by the contour detection algorithm as input and applies an ellipse fitting algorithm based on random sampling consensus (RANSAC) or least squares. The algorithm can effectively resist the interference of a small number of noise points generated by edge detection and fit an ellipse that best represents the pupil contour. The specific process of the ellipse fitting algorithm of random sampling consensus (RANSAC) is to randomly extract a minimum number of points from the edge point set (5 points are required to fit an ellipse) and fit a candidate ellipse. Then calculate the distance from all edge points to this candidate ellipse, and count the number of inner points whose distance is less than a certain threshold. Repeat this process multiple times, and select the candidate ellipse with the most inner points as the best fitting result. Once the best fitting ellipse is determined, the geometric center coordinates of the ellipse are accurately calculated and determined as the center coordinates of the final geometric center of the pupil.

[0080] Step S20 , performing data fusion calculation on the head movement data and the eye movement data to calculate an absolute sight line vector pointing to the wearer's visual focus in a preset world coordinate system and compensating for head movement interference;

[0081] In one embodiment of the present invention, step S20 specifically includes:

[0082] The head motion data is processed by a posture solving algorithm to derive a rotation matrix or quaternion that describes the real-time rotation of the helmet body coordinate system relative to the preset world coordinate system;

[0083] The eye movement data is converted into a gaze unit direction vector in the helmet body coordinate system through a pre-calibrated mapping model;

[0084] Apply the rotation matrix or quaternion to the sight unit direction vector in the helmet body coordinate system, and through the coordinate system transformation operation, obtain the final absolute sight vector in the preset world coordinate system.

[0085] Specifically, for a clear description, we first define three key coordinate systems:

[0086] World Coordinate System: A fixed Cartesian coordinate system that does not move with the wearer. Typically, this system is established at system startup, using the helmet's initial posture as a reference. In an embodiment of the present invention, when the power is turned on, the helmet is placed on a level surface, and the world coordinate system is defined by the helmet's orientation at that moment. For example, the forward direction is the positive Y-axis, the right side is the positive X-axis, and vertically upward is the positive Z-axis. This coordinate system remains unchanged during system operation.

[0087] Helmet body coordinate system: A local coordinate system fixed to the helmet that moves with the head. Its origin can be set at the center of the head posture sensor or the geometric center of the helmet, and its coordinate axes are aligned with the sensitive axes of the head posture sensor.

[0088] Eye camera coordinate system: A coordinate system fixed to the eye tracking module camera, used to describe the movement of the eyeball.

[0089] Furthermore, the central processing unit acquires three-axis angular velocity data and three-axis acceleration data from the head posture sensor in real time and executes a posture solution algorithm based on a complementary filter algorithm, an extended Kalman filter (EKF), or a Madgwick / Mahony gradient descent algorithm. In this embodiment, the more optimized Madgwick algorithm is employed. This is a gradient descent-based, computationally efficient, and highly accurate AHRS (Attitude and Heading Reference System) algorithm. Specifically, the Madgwick algorithm maintains a quaternion representing the current head posture. This quaternion directly describes the rotation from the helmet body coordinate system to a preset world coordinate system. At each time step, the Madgwick algorithm first integrates the quaternion at the previous moment using only the angular velocity data to obtain a preliminary gyroscope-based posture estimate. The Madgwick algorithm then calculates a corrected gradient using acceleration data (used to correct pitch and roll angles and eliminate gyroscope drift) and magnetometer data (used to correct yaw angle if a nine-axis IMU is used). This modified gradient is then used to adjust the initial quaternion estimate using gradient descent, yielding a quaternion describing the current rotation of the helmet's coordinate system relative to the predefined world coordinate system. This quaternion contains all the information needed to rotate any vector in the helmet's coordinate system to the predefined world coordinate system.

[0090] Furthermore, the central processing unit (CPU) acquires eye movement data from the eye tracking module in real time and then uses a mapping model pre-calibrated during preliminary work. This mapping model is generated by executing a calibration procedure when the user first uses the headset. This calibration procedure guides the user to sequentially gaze at nine known points on the screen or in space. As the user gazes at each point, the corresponding eye movement data is recorded, and the known three-dimensional direction vector of that point relative to the user's eye in the helmet's coordinate system is calculated. Through the correspondence between these eye movement data and the direction vector of the known gaze point in the coordinate system of the helmet body, a mapping function is fitted using multivariate polynomial regression or training of a small neural network, where the output of the mapping function is a three-dimensional line of sight direction vector in the coordinate system of the helmet body. During normal use, the central processing unit inputs the real-time eye movement data into this stored mapping function and calculates a three-dimensional line of sight direction vector in the coordinate system of the helmet body. To facilitate subsequent calculations, the three-dimensional line of sight direction vector is normalized to obtain a line of sight unit direction vector, so that the real-time eye movement data can eventually be converted into a line of sight unit direction vector in the coordinate system of the helmet body.

[0091] Furthermore, the real-time rotation quaternion obtained in the above steps and the sight unit direction vector in the helmet body coordinate system are subjected to quaternion rotation vector calculation to realize data fusion, so that the sight unit direction vector is transformed from the helmet body coordinate system to the preset world coordinate system. The sight unit direction vector obtained by the calculation in the world coordinate system is the absolute sight vector that has completely compensated for the interference of head movement. At this time, no matter how the wearer's head turns, as long as his eyes continue to look at the same static point in the outside world, the direction of his absolute sight vector will remain unchanged, so that it will no longer be affected by the wearer's head posture, and only reflect the real direction of his sight in the fixed world. It provides a high-quality, uncoupled data foundation for subsequent eye movement pattern recognition and accurate and stable control of the light beam, thereby fundamentally solving the technical problem in the existing technology that the light beam cannot accurately match the real sight focus.

[0092] Step S30, performing a joint time domain and space domain analysis on the time series data stream consisting of the continuous absolute gaze vectors to identify the wearer's current eye movement behavior as one of at least two preset eye movement patterns;

[0093] In one embodiment of the present invention, the eye movement pattern includes at least a gaze pattern representing a stable visual focus and a glance pattern representing a rapid visual sweep. The above step S30 specifically includes:

[0094] Calculate the angular velocity of the absolute sight vector in real time based on the time series data stream;

[0095] comparing the calculated angular velocity with at least two preset velocity thresholds, the velocity thresholds comprising a higher glance velocity threshold and a lower gaze velocity threshold;

[0096] When the angular velocity is higher than the saccadic velocity threshold, the current eye movement behavior is identified as a saccadic pattern;

[0097] When the angular velocity is lower than the gaze velocity threshold, and its duration exceeds a preset gaze time threshold, and the spatial dispersion of the absolute gaze vector within its duration is lower than a preset spatial dispersion threshold, the current eye movement behavior is identified as a gaze pattern.

[0098] Specifically, the central processing unit maintains a sliding time window buffer to store all absolute sight vectors and their corresponding timestamps in the most recent period of time (for example, the last 500 milliseconds). For each newly arrived absolute sight vector, the central processing unit calculates its angular velocity with the absolute sight vector of the previous frame. Specifically, first, calculate the angle between the two unit vectors, the absolute sight vector of the previous frame and the current absolute sight vector, which can be calculated by the dot product formula. Then, the angle between the two unit vectors is divided by the time difference between the two frames (that is, the inverse of the sampling period) to obtain the instantaneous angular velocity.

[0099] Furthermore, the central processing unit performs a preliminary classification of the eye movement state of each frame based on the calculated instantaneous angular velocity and maintains a state timer and a temporary point set buffer. When the instantaneous angular velocity is determined to be higher than the saccade velocity threshold, the current eye movement behavior is identified as a saccade pattern. When the instantaneous angular velocity is determined to be lower than the gaze velocity threshold, the endpoint coordinates of the current absolute line of sight vector are added to the temporary point set buffer. At the same time, the gaze state timer is started or accumulated. At this time, the central processing unit continuously monitors the gaze state timer. When and only when the gaze state timer exceeds the preset gaze time threshold for the first time, the central processing unit triggers a spatial dispersion calculation for all points in the temporary point set buffer. There are multiple methods for calculating spatial dispersion. A simple and effective method is to calculate the maximum distance from all points in the point set added to the temporary point set buffer to their geometric center, or to calculate the maximum distance between any two points in the point set. In this case, for example, the maximum diameter (i.e., the maximum distance between two points) is taken as the spatial dispersion. If this calculated spatial dispersion is lower than a preset spatial dispersion threshold, this series of low-speed, concentrated eye movement behaviors is officially confirmed as a gaze pattern. At this time, the point set in the temporary point set buffer can also be packaged as a gaze event.

[0100] Therefore, in this embodiment of the present invention, angular velocity thresholds are first used to coarsely screen out high-speed glances and potential low-speed fixations. Then, for these potential fixations, a more rigorous fine-screening process combining duration and spatial concentration is used to ultimately confirm a valid fixation. This process ensures that only truly stable and focused visual lingering is recognized as fixation intent, providing a reliable decision-making basis for subsequent precise lighting control.

[0101] Furthermore, in one embodiment of the present invention, the eye movement mode also includes a smooth pursuit mode representing smooth visual pursuit of a dynamic target. The above step S30 specifically includes:

[0102] When the angular velocity is higher than the gaze velocity threshold and lower than the saccade velocity threshold, a motion prediction model is used to model the motion trajectory represented by the absolute gaze vector sequence in real time. At each time step, the theoretical position of the absolute gaze vector at the next moment is predicted based on the historical trajectory data.

[0103] Compare the predicted theoretical position with the actual position of the absolute sight vector at the next moment, and calculate the prediction error between the two;

[0104] The current eye movement behavior is identified as a smooth pursuit mode if and only if the angular velocity is continuously higher than the fixation velocity threshold and lower than the saccade velocity threshold within a preset minimum duration, and the calculated prediction error is continuously lower than a preset error threshold.

[0105] Specifically, when the angular velocity is higher than the gaze velocity threshold and lower than the scan velocity threshold, a motion prediction model is used to model the motion trajectory represented by the absolute line of sight vector sequence in real time, wherein the motion prediction model can be specifically implemented as a Kalman filter, an alpha-beta filter, or a simple linear regression model. In the embodiment of the present invention, a Kalman filter is specifically used to model the motion of the line of sight. The state vector of this filter can include position and velocity. When smooth tracking begins, the state of the Kalman filter is initialized with the first few points. In each subsequent frame, the Kalman filter will perform two steps: prediction and update. The prediction step is to predict the position where the line of sight should be at the current moment based on the state at the previous moment. The update step is to correct the prediction with the current actual observed line of sight position to obtain a more accurate state estimate at the current moment. Furthermore, a predicted position will be obtained in each prediction stage. The central processing unit calculates the angular distance between this predicted position and the actual observed position, that is, the prediction error. If the prediction error for multiple consecutive frames is less than a preset prediction error threshold (e.g., less than 1.5°), it indicates that the gaze movement is predictable and smooth, rather than random or jumpy. This is the fundamental difference between smooth pursuit and a series of small saccades. A series of dense, small saccades, while likely averaging at a moderate speed, is discontinuous and unpredictable, resulting in a large prediction error. By setting an error threshold, this type of pseudo-smooth pursuit can be effectively eliminated.

[0106] At this time, if the consistency of the motion trajectory is met, the duration exceeds the minimum duration of smooth tracking (for example, greater than 150ms), and the speed is within a specific range, the central processing unit will eventually formally confirm this series of "medium-speed, smooth, and predictable" eye movement behaviors as a smooth tracking mode, ensuring the high accuracy and robustness of the recognition results, thereby providing a reliable decision-making basis for realizing advanced dynamic target following lighting functions.

[0107] Step S40, executing a differentiated light source driving control strategy corresponding to the identified eye movement pattern according to the identified eye movement pattern;

[0108] In one embodiment of the present invention, the differentiated light source driving control strategy includes:

[0109] When the recognized eye movement pattern is a gaze pattern, a target lighting point is calculated based on a set of absolute sight line vectors constituting the recognized gaze pattern, and the rotatable light source on the stand lamp is driven to adjust its projection angle relative to the helmet in real time in a manner to offset the wearer's head movement, so as to project the light beam toward the target lighting point;

[0110] When the identified eye movement pattern is a scanning pattern, a light source angle suppression strategy is executed, which keeps the projection angle of the rotatable light source relative to the helmet stable during the scanning period.

[0111] In one embodiment of the present invention, the step of calculating a target lighting point based on a set of absolute sight line vectors constituting the gaze pattern identified this time includes:

[0112] Extracting candidate gaze clusters that are identified as gaze patterns and consist of a set of absolute gaze vectors;

[0113] The arithmetic mean or weighted mean of the coordinates of all vector endpoints in the candidate fixation cluster is calculated to determine the coordinates representing the geometric center of the candidate fixation cluster, and the coordinates of the geometric center of the candidate fixation cluster are determined as the target illumination point.

[0114] In one embodiment of the present invention, the step of driving the rotatable light source on the bracket lamp to adjust its projection angle relative to the helmet in real time in a manner to offset the wearer's head movement so as to project the light source beam toward the target illumination point includes:

[0115] The target lighting point in the preset world coordinate system is converted into the target projection vector in the helmet body coordinate system through an inverse coordinate system transformation based on real-time head motion data;

[0116] The target projection vector is decomposed into the target pitch angle and target yaw angle required for rotating the light source through the inverse kinematics model;

[0117] Comparing the target pitch angle and the target yaw angle with the current actual pitch angle and yaw angle in the head motion data to calculate the angle error;

[0118] A PID controller is used to take the angle error as input and continuously generate a control signal for driving a motor that can rotate the light source until the angle error converges to a preset range.

[0119] The central processing unit uses the inverse transformation of the current head posture rotation quaternion to reversely transform the coordinates of the target lighting point in the world coordinate system back to the helmet body coordinate system, obtaining the target projection vector. This is then decomposed into the target pitch and yaw angles that the two servo motors need to achieve using an inverse kinematics model. These two target angles are fed into two independent PID controllers as setpoints. The PID controllers then calculate the error based on the current actual angle feedback from the motor encoders and continuously output PWM signals to drive the motors, compensating for head movement in real time to ensure that the light beam is stably directed to the target lighting point.

[0120] Among them, when it is in the scanning mode, the central processing unit sends an angle suppression instruction to the motor driver, which enables the PID controller to freeze the target set point at the last valid angle before entering the scanning mode, so that the light source remains stable relative to the helmet and prevents the light beam from shaking rapidly with the line of sight.

[0121] Furthermore, in one embodiment of the present invention, corresponding to the above, when the eye movement pattern also includes a smooth tracking mode representing smooth visual pursuit of a dynamic target, the differentiated light source drive control strategy further includes:

[0122] When the identified eye movement pattern is a smooth tracking pattern, a dynamically moving target path is calculated based on the absolute line of sight vector sequence that constitutes the identified smooth tracking pattern, and the rotatable light source is driven to adjust its projection angle relative to the helmet in real time to smoothly move the light source beam along the target path at a matching speed.

[0123] In summary, the bracket lamp lighting control method applied to the helmet in the above embodiment of the present invention obtains head movement data and eye movement data in real time and synchronously, and performs data fusion calculation on the two to solve the absolute line of sight vector that compensates for the interference of head movement, thereby realizing the lighting control system to accurately lock the wearer's real visual focus, and solves the problem that the traditional headlamp light beam can only follow the physical orientation of the head but cannot be aligned with the wearer's real line of sight, ensuring that no matter how the wearer's head posture changes, the light beam can always be accurately projected to the target point he is observing; by performing a joint time domain and space domain analysis on the time series data stream composed of the absolute line of sight vector to identify the specific eye movement pattern and execute the corresponding differentiated light source drive control strategy, it realizes the intelligent judgment of the wearer's subjective intention, effectively avoids the frequent and ineffective shaking of the light beam caused by unconscious rapid scanning of the line of sight, and solves the problem of existing automatic head The problem of the lamp being unable to distinguish the user's intentions, easily causing visual interference and wearing discomfort; however, by calculating the precise target lighting point based on a set of absolute sight vectors constituting that gaze when a gaze pattern is identified, and driving the rotatable light source for directional projection, it ensures that the lighting energy is efficiently concentrated on the key area that the wearer is currently most interested in; and by implementing a light source angle suppression strategy when a scanning pattern is identified, the light source projection angle remains stable relative to the helmet during the scanning period, greatly enhancing the wearer's visual comfort and safety when observing a large environment or searching for a target; and by driving the rotatable light source to adjust the projection angle in a manner that offsets the wearer's head movement, the light beam is absolutely stable in the world coordinate system. When the wearer looks at a fixed point and turns their head at the same time, the light beam can always lock onto the sign without shaking with the head, ensuring continuous, uninterrupted and clear illumination of the target. This solves the problem of the existing inability to accurately determine user intentions and provide precise lighting.

[0124] Example 2

[0125] See also Figure 3 , is a schematic structural diagram of a lighting control device for a safety helmet bracket lamp provided in a second embodiment of the present invention. For ease of description, only portions related to the embodiment of the present invention are shown. The lighting control device for a safety helmet bracket lamp in the embodiment of the present invention includes:

[0126] A multi-source data acquisition module 11 is used to acquire head movement data representing the wearer's head posture in real time, and simultaneously acquire eye movement data representing the wearer's eye sight;

[0127] an absolute sight line vector calculation module 12 for performing data fusion calculation on the head movement data and the eye movement data to calculate an absolute sight line vector pointing to the wearer's visual focus in a preset world coordinate system and compensating for head movement interference;

[0128] an eye movement pattern recognition module 13 for performing a joint time-domain and space-domain analysis on a time-series data stream consisting of continuous absolute gaze vectors to identify the wearer's current eye movement behavior as one of at least two preset eye movement patterns, wherein the eye movement pattern includes at least a fixation pattern representing a stable visual focus and a saccade pattern representing a rapid visual sweep;

[0129] The lighting control module 14 is configured to execute a differentiated light source drive control strategy corresponding to the identified eye movement pattern according to the identified eye movement pattern, wherein the differentiated light source drive control strategy includes:

[0130] When the recognized eye movement pattern is a gaze pattern, a target lighting point is calculated based on a set of absolute sight line vectors constituting the recognized gaze pattern, and the rotatable light source on the stand lamp is driven to adjust its projection angle relative to the helmet in real time in a manner to offset the wearer's head movement, so as to project the light beam of the light source toward the target lighting point;

[0131] When the identified eye movement pattern is a scanning pattern, a light source angle suppression strategy is executed, wherein the light source angle suppression keeps the projection angle of the rotatable light source relative to the helmet stable during the scanning period.

[0132] Furthermore, in one embodiment of the present invention, the multi-source data acquisition module 11 includes:

[0133] an infrared light projection unit, configured to project infrared light toward the wearer's eye region using a near-infrared light source, so as to form at least two relatively fixed Purchinye light spots on the corneal surface;

[0134] An image capturing unit, configured to continuously capture a sequence of eye images including a pupil outline and the Purchinye light spot using an image sensor;

[0135] an image processing unit, configured to locate, in each frame of the eye image sequence, the geometric center of the pupil and the geometric centers of at least two Purchinye spots using an image processing algorithm;

[0136] The eye movement data generating unit is used to calculate the differential vector pointing from the geometric center of each Purchinye light spot to the geometric center of the pupil, and fuse the calculated differential vectors to generate the eye movement data.

[0137] Furthermore, in one embodiment of the present invention, the image processing unit includes:

[0138] The first image recognition subunit is used to identify and locate the geometric center of all Purchinye light spots in each frame of image by using a fast threshold segmentation algorithm;

[0139] a region delineation subunit, configured to delineate a region of interest containing only the pupil in each frame of image based on the geometric centers of the multiple located Purchinye light spots;

[0140] The second image recognition subunit is used to identify the outline of the pupil within the region of interest according to the outline detection algorithm, and fit the outline of the pupil by an ellipse fitting algorithm to determine the geometric center of the pupil.

[0141] Furthermore, in one embodiment of the present invention, the absolute sight vector solving module 12 includes:

[0142] A posture solving unit, configured to process the head motion data through a posture solving algorithm to derive a rotation matrix or quaternion describing the real-time rotation of the helmet body coordinate system relative to the preset world coordinate system;

[0143] A coordinate conversion unit, configured to convert the eye movement data into a sight unit direction vector in a helmet body coordinate system through a pre-calibrated mapping model;

[0144] The absolute sight line vector determination unit is used to apply the rotation matrix or quaternion to the sight line unit direction vector in the helmet body coordinate system, and obtain the final absolute sight line vector in the preset world coordinate system through coordinate system transformation operation.

[0145] Furthermore, in one embodiment of the present invention, the eye movement pattern recognition module 13 includes:

[0146] a data calculation unit, configured to calculate the angular velocity of the absolute sight line vector in real time based on the time series data stream;

[0147] a data comparison unit, configured to compare the calculated angular velocity with at least two preset velocity thresholds, wherein the velocity thresholds include a higher scanning velocity threshold and a lower gaze velocity threshold;

[0148] a scanning pattern recognition unit, configured to recognize the current eye movement behavior as a scanning pattern when the angular velocity is higher than the scanning velocity threshold;

[0149] A gaze pattern recognition unit is used to identify the current eye movement behavior as a gaze pattern when the angular velocity is lower than the gaze velocity threshold, its duration exceeds a preset gaze time threshold, and the spatial discreteness of the absolute line of sight vector within its duration is lower than a preset spatial dispersion threshold.

[0150] Furthermore, in one embodiment of the present invention, the lighting control module 14 includes:

[0151] a candidate gaze cluster extraction unit, configured to extract a candidate gaze cluster that is identified as a gaze pattern and is composed of a set of absolute gaze vectors;

[0152] The target lighting point determination unit is used to calculate the arithmetic mean or weighted mean of the coordinates of all vector endpoints in the candidate fixation cluster, determine the coordinates representing the geometric center of the candidate fixation cluster, and determine the coordinates of the geometric center of the candidate fixation cluster as the target lighting point.

[0153] Furthermore, in one embodiment of the present invention, the lighting control module 14 includes:

[0154] A coordinate inverse conversion unit is used to convert the target illumination point in the preset world coordinate system into a target projection vector in the helmet body coordinate system through an inverse transformation of the coordinate system based on real-time head motion data;

[0155] a vector decomposition unit, configured to decompose the target projection vector into a target pitch angle and a target yaw angle required by the rotatable light source through an inverse kinematics model;

[0156] an angle error calculation unit, configured to compare the target pitch angle and target yaw angle with the current actual pitch angle and yaw angle in the head motion data to calculate an angle error;

[0157] The control signal generating unit is configured to use a PID controller to take the angle error as input and continuously generate a control signal for driving the motor of the rotatable light source until the angle error converges to a preset range.

[0158] Furthermore, in one embodiment of the present invention, the eye movement pattern identified by the eye movement pattern recognition module further includes a smooth pursuit pattern representing smooth visual pursuit of a dynamic target, and the differentiated light source drive control strategy in the lighting control module further includes:

[0159] When the identified eye movement pattern is a smooth tracking pattern, a dynamically moving target path is calculated based on the absolute line of sight vector sequence that constitutes the identified smooth tracking pattern, and the rotatable light source is driven to adjust its projection angle relative to the helmet in real time to smoothly move the light source beam along the target path at a matching speed.

[0160] The implementation principle and technical effects of the bracket lamp lighting control device for a safety helmet provided in the embodiment of the present invention are the same as those of the aforementioned method embodiment. For the sake of brief description, for matters not mentioned in the device embodiment, reference can be made to the corresponding contents in the aforementioned method embodiment.

[0161] Example 3

[0162] Another aspect of the present invention also provides a lighting control system for a safety helmet bracket lamp, see Figure 4 , which is shown as a third embodiment of the present invention, a bracket lamp lighting control system applied to a helmet includes:

[0163] Safety helmet 10;

[0164] A head posture sensor 20 is mounted on the helmet 10 and is used to obtain head movement data of the wearer;

[0165] An eye tracking module 30 , integrated into the helmet 10 or its mounting accessories, for acquiring eye movement data of the wearer;

[0166] A bracket lamp 40 is mounted on the helmet 10 and includes a bracket and a rotatable light source;

[0167] The central processing unit 50 is electrically connected to the head posture sensor 20, the eye tracking module 30 and the bracket light 40. The central processing unit 50 is configured to execute the bracket light lighting control method applied to the helmet as described in the above embodiment.

[0168] Wherein, in an embodiment of the present invention, the helmet is a helmet that meets industrial safety standards. Wherein, the head posture sensor adopts a nine-axis inertial measurement unit (IMU), which integrates a three-axis accelerometer, a three-axis gyroscope and a three-axis magnetometer, and is firmly mounted at the top center of the helmet. The head posture sensor is used to obtain the wearer's head pitch angle, yaw angle, roll angle and its angular velocity, angular acceleration and other head motion data in real time. Wherein, the eye tracking module is integrated into the forehead position of the helmet lining or set on the mounting accessories of the helmet (such as installed under the brim of the helmet through an adjustable bracket), wherein the eye tracking module includes a global shutter near-infrared (NIR) camera and two 850nm infrared LED light sources distributed on both sides of the camera in an asymmetrical layout, wherein the near-infrared camera is sensitive to infrared light with a wavelength of 850nm and is equipped with an infrared passband filter, and the eye tracking module is used to obtain the wearer's eye movement data. The bracket light is a lighting assembly mounted on the forehead of a hardhat. It includes a two-dimensional gimbal bracket driven by two high-precision, low-noise micro-servo motors and a high-brightness LED light source mounted on the gimbal bracket. The two micro-servo motors respectively control the pitch (up and down) and yaw (left and right) rotation of the LED light source, and each micro-digital servo motor is equipped with a high-resolution magnetic encoder for real-time angle feedback. The central processing unit, a high-performance microcontroller, is mounted in the battery compartment at the rear of the hardhat. The central processing unit communicates with the head posture sensor via the I2C bus and with the eye tracking module via a USB or SPI interface. It also outputs PWM signals to control the two micro-servo motors of the bracket light. The central processing unit is configured to execute the bracket light lighting control method for hardhats as described in the aforementioned embodiments.

[0169] In summary, the bracket lamp lighting control system applied to the helmet in the above embodiment of the present invention obtains head movement data and eye movement data in real time and synchronously, and performs data fusion calculation on the two to solve the absolute line of sight vector that compensates for the interference of head movement, thereby realizing the lighting control system to accurately lock the wearer's real visual focus, solving the problem that the traditional headlamp light beam can only follow the physical orientation of the head but cannot be aligned with the wearer's real line of sight, and ensuring that no matter how the wearer's head posture changes, the light beam can always be accurately projected to the target point he is observing; by performing a joint time domain and space domain analysis on the time series data stream composed of the absolute line of sight vector to identify the specific eye movement pattern and execute the corresponding differentiated light source drive control strategy, it realizes the intelligent judgment of the wearer's subjective intention, effectively avoids the frequent and ineffective shaking of the light beam caused by unconscious rapid scanning of the line of sight, and solves the problem of existing automatic head The problem of the lamp being unable to distinguish the user's intentions, easily causing visual interference and wearing discomfort; however, by calculating the precise target lighting point based on a set of absolute sight vectors constituting that gaze when a gaze pattern is identified, and driving the rotatable light source for directional projection, it ensures that the lighting energy is efficiently concentrated on the key area that the wearer is currently most interested in; and by implementing a light source angle suppression strategy when a scanning pattern is identified, the light source projection angle remains stable relative to the helmet during the scanning period, greatly enhancing the wearer's visual comfort and safety when observing a large environment or searching for a target; and by driving the rotatable light source to adjust the projection angle in a manner that offsets the wearer's head movement, the light beam is absolutely stable in the world coordinate system. When the wearer looks at a fixed point and turns their head at the same time, the light beam can always lock onto the sign without shaking with the head, ensuring continuous, uninterrupted and clear illumination of the target. This solves the problem of the existing inability to accurately determine user intentions and provide precise lighting.

[0170] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" means that a specific feature, structure, material, or characteristic described in conjunction with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0171] The above-described embodiments merely illustrate several embodiments of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art would be able to make various modifications and improvements without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

Claims

1. A lighting control method for a safety helmet bracket lamp, characterized in that: The method comprises: Acquire head movement data representing the wearer's head posture in real time, and simultaneously acquire eye movement data representing the wearer's eye gaze; Performing data fusion calculation on the head movement data and the eye movement data to calculate an absolute sight line vector pointing to the wearer's visual focus in a preset world coordinate system and compensating for head movement interference; Performing a joint time-domain and space-domain analysis on a time-series data stream consisting of continuous absolute gaze vectors to identify the wearer's current eye movement behavior as one of at least two preset eye movement patterns, wherein the eye movement pattern includes at least a fixation pattern representing a stable visual focus and a saccade pattern representing a rapid visual sweep; According to the identified eye movement pattern, a differentiated light source drive control strategy corresponding to the identified eye movement pattern is executed, wherein the differentiated light source drive control strategy includes: When the recognized eye movement pattern is a gaze pattern, a target lighting point is calculated based on a set of absolute sight line vectors constituting the recognized gaze pattern, and the rotatable light source on the stand lamp is driven to adjust its projection angle relative to the helmet in real time in a manner to offset the wearer's head movement, so as to project the light beam of the light source toward the target lighting point; When the identified eye movement pattern is a scanning pattern, a light source angle suppression strategy is executed, wherein the light source angle suppression keeps the projection angle of the rotatable light source relative to the helmet stable during the scanning period.

2. The method for controlling the lighting of a safety helmet bracket lamp according to claim 1, wherein: The step of synchronously acquiring eye movement data representing the wearer's eye sight includes: Using a near-infrared light source to project infrared light toward the wearer's eye area, so as to form at least two relatively fixed Purchinye light spots on the corneal surface; Continuously capturing a sequence of eye images including the pupil outline and the Purchinye light spot using an image sensor; In each frame of the eye image sequence, the geometric center of the pupil and the geometric centers of at least two Purchinye spots are located by an image processing algorithm; A differential vector pointing from the geometric center of each Purchinye light spot to the geometric center of the pupil is calculated, and the calculated differential vectors are fused to generate the eye movement data.

3. The lighting control method for a safety helmet bracket lamp according to claim 2, characterized in that: The step of locating the geometric center of the pupil and the geometric centers of at least two Purchinye light spots by using an image processing algorithm comprises: In each frame of the image, the geometric center of all Purchinye spots is identified and located by a fast threshold segmentation algorithm; Based on the geometric centers of the multiple located Purchinye light spots, a region of interest containing only the pupil is delineated in each frame image; In the region of interest, the pupil contour is identified according to a contour detection algorithm, and the pupil contour is fitted using an ellipse fitting algorithm to determine the geometric center of the pupil.

4. The lighting control method for a safety helmet bracket lamp according to claim 1, characterized in that: The step of performing data fusion calculation on the head movement data and the eye movement data to calculate an absolute sight line vector pointing to the wearer's visual focus in a preset world coordinate system and compensating for head movement interference includes: Processing the head motion data through a posture solving algorithm to derive a rotation matrix or quaternion describing the real-time rotation of the helmet body coordinate system relative to the preset world coordinate system; Converting the eye movement data into a sight unit direction vector in the helmet body coordinate system through a pre-calibrated mapping model; The rotation matrix or quaternion is applied to the sight unit direction vector in the helmet body coordinate system, and the absolute sight vector in the preset world coordinate system is finally obtained through coordinate system transformation operation.

5. The lighting control method for a safety helmet bracket lamp according to claim 1, characterized in that: The step of performing a joint time domain and space domain analysis on the time series data stream consisting of the continuous absolute gaze vectors to identify the wearer's current eye movement behavior as one of at least two preset eye movement patterns includes: Calculating the angular velocity of the absolute sight line vector in real time based on the time series data stream; comparing the calculated angular velocity with at least two preset velocity thresholds, the velocity thresholds comprising a higher glance velocity threshold and a lower gaze velocity threshold; When the angular velocity is higher than the saccadic velocity threshold, the current eye movement behavior is identified as a saccadic mode; When the angular velocity is lower than the gaze velocity threshold, and its duration exceeds a preset gaze time threshold, and the spatial discreteness of the absolute gaze vector within its duration is lower than a preset spatial dispersion threshold, the current eye movement behavior is identified as a gaze pattern.

6. The lighting control method for a safety helmet bracket lamp according to claim 1, characterized in that: The step of calculating a target lighting point based on a set of absolute sight line vectors constituting the gaze pattern identified this time comprises: Extracting candidate gaze clusters that are identified as gaze patterns and consist of a set of absolute gaze vectors; The arithmetic mean or weighted mean of the coordinates of all vector endpoints in the candidate fixation cluster is calculated to determine the coordinates representing the geometric center of the candidate fixation cluster, and the coordinates of the geometric center of the candidate fixation cluster are determined as the target illumination point.

7. The lighting control method for a safety helmet bracket lamp according to claim 1, characterized in that: The step of driving the rotatable light source on the bracket lamp to adjust its projection angle relative to the helmet in real time in a manner to offset the wearer's head movement so as to project the light source beam toward the target lighting point includes: The target lighting point in the preset world coordinate system is converted into the target projection vector in the helmet body coordinate system through an inverse coordinate system transformation based on real-time head motion data; Decomposing the target projection vector into a target pitch angle and a target yaw angle required by the rotatable light source through an inverse kinematics model; comparing the target pitch angle and target yaw angle with the current actual pitch angle and yaw angle in the head motion data to calculate an angle error; A PID controller is used to take the angle error as input and continuously generate a control signal for driving a motor of the rotatable light source until the angle error converges to a preset range.

8. The lighting control method for a safety helmet bracket lamp according to claim 1, characterized in that: The eye movement mode also includes a smooth tracking mode representing smooth visual pursuit of a dynamic target, and the differentiated light source drive control strategy also includes: When the identified eye movement pattern is a smooth tracking pattern, a dynamically moving target path is calculated based on the absolute line of sight vector sequence that constitutes the identified smooth tracking pattern, and the rotatable light source is driven to adjust its projection angle relative to the helmet in real time to smoothly move the light source beam along the target path at a matching speed.

9. A lighting control device for a safety helmet bracket lamp, characterized in that: The device comprises: A multi-source data acquisition module is used to acquire head movement data representing the wearer's head posture in real time, and simultaneously acquire eye movement data representing the wearer's eye sight; an absolute sight line vector calculation module, configured to perform data fusion calculation on the head movement data and the eye movement data to calculate an absolute sight line vector pointing to the wearer's visual focus in a preset world coordinate system and compensating for head movement interference; an eye movement pattern recognition module, configured to perform a joint temporal and spatial analysis on a time-series data stream consisting of continuous absolute gaze vectors to identify the wearer's current eye movement behavior as one of at least two preset eye movement patterns, wherein the eye movement pattern includes at least a gaze pattern representing a stable visual focus and a saccade pattern representing a rapid visual sweep; The lighting control module is configured to execute a differentiated light source drive control strategy corresponding to the identified eye movement pattern according to the identified eye movement pattern, wherein the differentiated light source drive control strategy includes: When the recognized eye movement pattern is a gaze pattern, a target lighting point is calculated based on a set of absolute sight line vectors constituting the recognized gaze pattern, and the rotatable light source on the stand lamp is driven to adjust its projection angle relative to the helmet in real time in a manner to offset the wearer's head movement, so as to project the light beam of the light source toward the target lighting point; When the identified eye movement pattern is a scanning pattern, a light source angle suppression strategy is executed, wherein the light source angle suppression keeps the projection angle of the rotatable light source relative to the helmet stable during the scanning period.

10. A lighting control system for a safety helmet bracket lamp, characterized in that: The system comprises: helmet; A head posture sensor, mounted on the helmet, for acquiring head movement data of the wearer; an eye tracking module, integrated into the helmet or its mounting accessories, for acquiring eye movement data of the wearer; a bracket lamp, mounted on the safety helmet, comprising a bracket and a rotatable light source; A central processing unit is electrically connected to the head posture sensor, the eye tracking module and the bracket light, and the central processing unit is configured to execute the bracket light lighting control method applied to a safety helmet as described in any one of claims 1 to 8.

Images