A pointing interaction method and device based on polarization vision anchor

Abstract

The application discloses a pointing interaction method and device based on polarization vision anchor, and belongs to the technical field of human-computer interaction. The method comprises the following steps: acquiring image data of a display scene under at least two different polarization analysis states, calculating a polarization response feature and extracting a display area, and establishing a visual pointing coordinate; collecting angular motion data at a frequency higher than a visual observation frequency, and performing high-frequency pointing prediction between adjacent visual observation periods; dynamically updating mapping parameters of the angular motion quantity to the screen coordinate displacement by using a continuous visual observation coordinate change amount and an angular motion integral amount in a corresponding time period; and outputting continuous pointing coordinates and an interaction instruction according to the updated mapping parameters. The application does not need to modify a display end, has low computing power requirement, can effectively inhibit drift, adapts to user distance and posture changes, and provides a low-delay and high-fluency pointing interaction experience.

Description

Technical Field

[0001] This invention relates to the field of human-computer interaction technology, and in particular to a pointing interaction method and device based on polarization visual anchoring. Background Technology

[0002] In long-distance interactive scenarios such as smart TVs, conference screens, educational display terminals, and projection devices, users frequently need to perform tasks such as icon selection, text input, menu operation, page scrolling, and presentation control. Traditional directional remote controls achieve selection by moving the focus item by item, which is inefficient. While touchpad remote controls improve the swipe input experience, there is a lack of intuitive spatial correspondence between the touch area and the display screen. Therefore, "pointing" interaction technology, similar to a laser pointer or mouse, has become a research hotspot in the field of human-computer interaction.

[0003] Currently, existing pointing interaction technologies can be mainly divided into the following categories: The first type is the pointing control scheme based on inertial sensors. This scheme uses inertial sensors such as gyroscopes and accelerometers to collect the attitude changes of the handheld device, and estimates the cursor movement through angular velocity integration or attitude calculation. It has low hardware cost and fast response speed, but it is essentially a relative motion measurement and lacks an absolute spatial reference from the display screen. The gyroscope output will accumulate drift after integration, and the position of the display screen in the camera's field of view or in real space cannot be directly determined. When the user changes their sitting posture, grip angle, distance, or picks up the remote control again, the cursor position may deviate significantly from the actual pointing direction.

[0004] The second category is conventional visual recognition solutions based on cameras. This approach captures images of the display screen and its surrounding environment using a camera, identifies screen areas based on brightness, color, edges, corners, contours, or deep learning models, and then maps the pointing location to screen coordinates through perspective transformation. While this method provides an absolute visual reference, it is easily affected by changes in screen content, ambient light, reflections, low light, background complexity, and computational limitations. If relying solely on brightness, color, or edge features for screen recognition, changes in screen content, screen blackouts, similar background brightness, glass reflections, or strong light can all lead to recognition failures. Using deep learning models to improve robustness requires a large amount of training data and significant computational resources.

[0005] The third type is a positioning solution based on infrared, laser, optical markers, or active signals. This solution places infrared emitters, laser markers, or optical marker patterns around the display screen or on the display terminal side, and the handheld device achieves positioning by recognizing the external markers. This solution provides clear positioning, but requires modification of the display end or additional hardware, and is not suitable for direct compatibility with a large number of existing display devices.

[0006] The fourth category is based on wireless ranging or spatial positioning. This approach establishes the spatial relationship between the handheld device and the display device through methods such as UWB, Bluetooth ranging, ultrasound, magnetic positioning, or multi-base station positioning. It typically requires both the display and remote control terminals to have dedicated communication or positioning hardware, resulting in higher system complexity and cost.

[0007] In summary, existing directional interaction solutions struggle to simultaneously achieve absolute spatial reference, low-latency response, and anti-drift performance without modifying the display or relying on high-performance visual models. Therefore, a directional interaction technology that can address these requirements is urgently needed. Summary of the Invention

[0008] This invention provides a pointing interaction method and device based on polarization visual anchoring, aiming to solve the contradiction in existing pointing interaction technologies where low-cost solutions lack absolute spatial reference and are prone to drift, while high-precision solutions rely on display modifications or high-computing-power visual models.

[0009] Specifically, this invention aims to solve the following technical problems: how to provide a stable visual anchor point for a handheld pointing device without modifying the display end or setting active markers; how to avoid excessive reliance on screen content and ambient light in ordinary visual recognition, thereby reducing computing power requirements; how to obtain high-frequency, smooth, and low-latency pointing output even when the visual observation frequency is low; and how to dynamically calibrate the mapping relationship between angular motion and screen coordinate displacement based on changes in user distance and grip posture, thereby suppressing and reducing the cumulative drift and proportional mismatch of pure inertial schemes.

[0010] To achieve the above-mentioned objectives, the present invention provides the following technical solution: In a first aspect, the present invention provides a pointing interaction method based on polarization visual anchoring, comprising the following steps: S1. Acquire image data of the displayed scene under at least two different polarization analysis states using a handheld pointing device; S2. Based on the image data from the at least two different polarization analysis states, calculate the polarization response characteristics used to distinguish the display area from the environmental background; S3. Perform dynamic threshold segmentation based on the polarization response feature map, and filter out candidate display areas in combination with geometric constraints; S4. Establish a mapping relationship between the image coordinate system and the display area coordinate system based on the selected display area to obtain the visual pointing coordinates; S5. Collect angular motion data at a sampling frequency higher than the visual observation frequency, and convert the angular motion change into the display area coordinate change between adjacent visual observation cycles according to the current mapping parameters, and output high-frequency predicted pointing coordinates. S6. When the change in visual pointing coordinates obtained at two consecutive visual observation times meets the preset confidence condition, the mapping parameter between the change in visual coordinates and the integral of angular motion within the corresponding time period is dynamically updated using the change in visual coordinates and the integral of angular motion within the corresponding time period. S7. Based on the updated mapping parameters, convert the angular motion data into continuous pointing coordinates, and encapsulate them into interactive commands for output to the display terminal.

[0011] Furthermore, the polarization response feature is at least one of the following: normalized differential feature, brightness difference, brightness ratio, polarization degree approximation, or response consistency feature under multiple polarization states; when the normalized differential feature is used, its calculation formula is:

[0012] in , These are the brightness values ​​under two different polarization analysis states. To prevent positive numbers with a denominator of zero.

[0013] Furthermore, the geometric constraints include one or more combinations of the following: connected region area range, boundary continuity, quadrilateral shape fitting error of candidate regions, aspect ratio range, number of corner points, boundary straightness, and temporal consistency with the display area position of the previous frame; the determination of the display area candidates adopts a combined judgment method of "polarization response difference + geometric constraints + temporal confidence".

[0014] Furthermore, the dynamic updating of mapping parameters employs a recursive filtering method:

[0015] in Changes in current visual coordinates Integral quantity of angular motion in the same time period Calculations show that The updated weights are adaptively adjusted based on visual confidence, angular motion amplitude, or angular velocity noise level, and ; Wherein, the angular motion integral quantity The mapping parameters can be a vector containing changes in yaw and pitch angles. It can be a scalar scaling factor, a two-dimensional matrix, a local linear mapping matrix, or a nonlinear correction parameter; When using a two-dimensional angular motion integral vector It can be obtained by using least squares, recursive least squares, or local fitting with regularization terms from the changes in visual coordinates and integrals of angular motion over multiple consecutive visual observation cycles.

[0016] Furthermore, it also includes confidence assessment and anomaly handling steps: A comprehensive confidence score is calculated for each visual observation result. The comprehensive confidence score is determined by at least two of the following factors: polarization response difference intensity, candidate region area, boundary straightness, corner stability, aspect ratio rationality, and consistency with historical position. When the overall confidence level is higher than the first threshold, the visual observation results are used to update the current pointing coordinates and mapping parameters; When the overall confidence level is lower than the second threshold, the mapping parameter update is paused, and only the previous effective mapping parameter is used for short-term angular motion prediction. Boundary constraints or velocity damping are applied to the predicted coordinates until the visual observation returns to stability. When the overall confidence level is between the first and second thresholds, reduce the weight of visual observations on the update of mapping parameters, or use only visual observations for slow correction.

[0017] Secondly, the present invention provides a pointing interaction device based on polarization visual anchoring, comprising: The polarization image acquisition module is used to acquire image data of the display scene under at least two different polarization analysis states; The display area recognition module is used to extract the display area based on polarization response differences and geometric constraints. The coordinate mapping module is used to establish a screen coordinate system or a pointing coordinate mapping relationship based on the observation results of the display area. An angular motion acquisition module is used to acquire the angular velocity or angular motion change of a handheld device. The dynamic calibration module is used to update the mapping parameters from angular motion to screen coordinate displacement based on continuous visual observation and synchronous angular motion integral. The interactive output module is used to output cursor coordinates, click, drag, or scroll interaction commands.

[0018] Furthermore, the polarization image acquisition module has any of the following structures: Two imaging channels with fixed polarization analysis directions, the fields of view of the two channels are basically overlapping, and polarization analyzers with different polarization directions are set respectively; An integrated polarization image sensor acquires pixel responses from multiple different polarization analysis directions on the same imaging chip; A single imaging channel, in conjunction with an electronically controlled polarization switching element, can be switched to at least two different polarization analysis states in time division, and an angular motion sensor is used to compensate for attitude changes between time division images.

[0019] Furthermore, the angular motion acquisition module is a gyroscope, and the device does not use the integral result of the accelerometer as a necessary input for the calculation of the pointing coordinates; the accelerometer is only used as an optional module for device wake-up, static state detection, power consumption control, or coarse attitude judgment.

[0020] Furthermore, it also includes a power consumption control module: When the pointing function is not enabled, the device is stationary, or there is no effective angular movement for a long time, the sampling frequency of the polarization image acquisition module is reduced or it is put into a sleep state. When the pointing function is detected to be enabled, the device is picked up, or the angular motion amplitude exceeds the preset threshold, the image acquisition frequency is restored or increased. During the pointing function, if the cursor does not move effectively or the visual confidence level is higher than the preset threshold, the visual observation frequency is dynamically reduced; if the angular motion amplitude increases, the visual confidence level decreases, or the cursor approaches the boundary of the display area, the visual observation frequency is temporarily increased.

[0021] Furthermore, the instruction format output by the interactive output module is Bluetooth HID instruction, USB HID instruction, private wireless protocol instruction, or application layer communication protocol instruction; the device is a handheld pointing device, and its form includes, but is not limited to, a TV remote control, a conference screen pointing device, an educational whiteboard controller, a presentation controller, or a smart home pointing controller.

[0022] Compared with the prior art, the present invention has the following advantages and beneficial effects: 1. Obtain visual anchor points without modifying the display. This invention utilizes the difference in polarization response characteristics between the display area and the environmental background for identification. It does not require the display to show special patterns, nor does it require the installation of infrared lights, markers, base stations, or other active hardware around the display screen. It is suitable for scenarios with detectable polarization response differences in existing display devices.

[0023] 2. Reduces the computational requirements of ordinary visual recognition. Compared with visual recognition methods that rely solely on color, brightness, texture, or deep learning models, this invention enhances the separability of the display area and background through polarization response features, enabling display area recognition to be completed using lightweight image processing steps such as difference, normalization, threshold segmentation, connected components, and geometric constraints.

[0024] 3. Improve the drift problem of pure inertial pointing. Visual observation results provide a low-frequency spatial correction reference for angular motion prediction, so that the system does not rely entirely on gyroscope integration; when the angular motion sensor has zero bias, proportional error or long-term integration error, it can be corrected by continuous visual observation results.

[0025] 4. Improve the continuity of interaction under low-frequency visual conditions. Visual recognition does not need to operate at the same frequency as the cursor output. High-frequency angular motion data can compensate for cursor movement between visual frames, thereby maintaining a high pointing coordinate output frequency while reducing image acquisition and processing power consumption.

[0026] 5. Adapts to changes in user distance and grip. By dynamically updating mapping parameters through continuous visual coordinate changes and angular motion integrals over the same time period, this invention can continuously correct the screen coordinate displacement relationship corresponding to angular motion when the user's posture, distance, grip angle, and operation range change, reducing cursor speed inconsistencies and unresponsiveness caused by fixed-ratio mapping.

[0027] 6. Improve system robustness. This invention employs a combination of polarization response differences, geometric constraints, and temporal confidence levels for judgment, rather than relying solely on single brightness or polarization differences. This reduces the impact of ambient light, local reflections, and changes in screen content on the recognition results. Attached Figure Description

[0028] Figure 1 Schematic diagram of the overall structure of a pointing interaction system based on polarization visual anchoring; Figure 2 Schematic diagram of the preferred structure of the polarization image acquisition module; Figure 3 Flowchart of a pointing interaction method based on polarization visual anchoring; Figure 4 : Time series relationship between low-frequency visual observation and high-frequency angular motion prediction; Figure 5 Schematic diagram of dynamic calibration of angular motion and coordinate displacement of the display area; Figure 6 : Schematic diagram of confidence level processing under abnormal visual observation. Detailed Implementation

[0029] The present invention will be further described in detail below with reference to experimental examples and specific embodiments. However, this should not be construed as limiting the scope of the above-mentioned subject matter of the present invention to the following embodiments; all technologies implemented based on the content of the present invention fall within the scope of the present invention.

[0030] This invention comprises five parts: polarization visual anchoring, display area coordinate establishment, high-frequency prediction of angular motion, dynamic mapping calibration, and interactive command output. The overall system structure and main module connections of this invention are as follows: Figure 1 As shown.

[0031] (I) Overall Technical Principles The display area and the ambient background typically exhibit detectable differences in polarization response characteristics. Here, "polarization response characteristics" are not limited to the display screen necessarily emitting linearly polarized light in a fixed direction, but rather refer to the brightness response, response difference, response ratio, degree of polarization, or polarization direction correlation characteristics of the display area compared to the ambient background under at least two different polarization analysis states. By acquiring images or optical responses under different polarization analysis states, more stable and easily segmented visual observation results of the display area relative to the ambient background can be obtained.

[0032] This invention does not rely on displaying specific patterns on the screen, nor does it require the display end to have infrared light strips, QR codes, special borders, or active emission modules. Polarization vision only undertakes the task of low-frequency, absolute or quasi-absolute observation of the display area; the angular motion sensor undertakes the task of high-frequency, low-latency continuous pointing prediction; the two establish a closed loop through dynamic mapping calibration.

[0033] (II) System Composition The system composition and the input / output relationships between its modules can be combined Figure 1 Understanding; the preferred hardware structure of the polarization image acquisition module can be found in [reference needed]. Figure 2 The system's composition and functions are shown in the table below:

[0034] (III) Methods and Procedures The overall flow of method steps S1 to S7 of the present invention is as follows: Figure 3 As shown, Figure 3 The steps below correspond one-to-one with the numbers below.

[0035] Step S1: Polarization vision data acquisition.

[0036] A handheld pointing device acquires image data of the displayed scene under at least two different polarization analysis states. Preferably, a dual-imaging-channel structure can be used, with polarization analyzers of different polarization directions placed in front of the two imaging channels, and the two imaging channels acquiring images within the same time window; alternatively, an integrated polarization image sensor can be used to acquire pixel responses in multiple polarization directions within the same frame. To ensure the reliability of the difference calculation, exposure consistency correction, lens distortion correction, image registration, and brightness normalization processing are preferably performed. The structural relationship of the dual-imaging-channel polarization image acquisition module can be found in [reference needed]. Figure 2 .

[0037] Step S2: Calculation of polarization response characteristics.

[0038] Based on image data under different polarization analysis states, polarization response features used to distinguish the display area from the environmental background are calculated. These polarization response features can be brightness difference, normalized difference, brightness ratio, polarization degree approximation, or response consistency features under multiple polarization states. In a preferred embodiment, normalized difference features are used:

[0039] in, and These are the brightness values ​​under two polarization analysis states, respectively, with ε being a positive number to prevent the denominator from being too small.

[0040] Step S3: Display region candidate extraction.

[0041] Dynamic threshold segmentation is performed based on polarization response feature maps to obtain candidate regions. Then, geometric constraints such as connected region area, boundary continuity, approximate quadrilateral shape, aspect ratio range, number of corner points, boundary straightness, and temporal stability are combined to determine candidate display regions. To avoid local polarization interference caused by environmental reflections, this invention does not rely solely on polarization difference as the sole criterion, but instead employs a combined determination method of "polarization response difference + display region geometric constraints + temporal confidence." This step is consistent with... Figure 3 S3 in the diagram corresponds to the candidate extraction process of the display area.

[0042] Step S4: Display the area coordinate system or point to the coordinate mapping establishment.

[0043] After identifying the boundaries or corners of the display area, a mapping relationship is established between the image coordinate system and the display area coordinate system. Preferably, when the four corners of the display area are available, a perspective transformation is established using a homography matrix to map the pointing center, optical axis intersection, or preset reference point in the camera image to the display area coordinate system, thus obtaining the visual pointing coordinates. When the confidence level of the displayed area observation is insufficient, the visual result is not forced to directly replace the current cursor coordinates. Instead, its weight is reduced, or it is only used for boundary constraints, slow drift correction, and mapping parameter updates. The processing logic for insufficient visual observation confidence can be found in [link to relevant documentation]. Figure 6 .

[0044] Step S5: High-frequency prediction of angular motion.

[0045] The handheld pointing device acquires angular motion data via an angular velocity sensor at a sampling frequency higher than the visual observation frequency. After zero-bias estimation, low-pass filtering, or stationary state correction, the angular change is obtained by integrating the angular motion data over adjacent visual observation cycles. Based on the current mapping parameters , the change in angle Convert to display area coordinate change and update the predicted pointing coordinates. The temporal relationship between low-frequency visual observation and high-frequency angular motion prediction can be found in [reference needed]. Figure 4 .

[0046] Step S6: Dynamic mapping calibration.

[0047] Obtain the visual pointing coordinates at the k-th visual observation time In the Visual pointing coordinates were obtained at each visual observation moment. Calculate the change in visual coordinates between the two. Simultaneously, the angular velocity data within the same time period are integrated to obtain the change in angular motion. .when and The visual confidence scores all meet the preset conditions, and When not close to zero, utilize and Update the mapping parameters between angular motion and display area coordinate displacement. The dynamic mapping calibration relationship can be found in [reference needed]. Figure 5 .

[0048] In a preferred embodiment, a recursive update method is used: ,in Depend on and Calculations show that To update the weights. The updated weights can be adaptively adjusted based on visual confidence, motion amplitude, candidate region stability, and angular velocity noise level.

[0049] Step S7: Point to coordinates and output interactive commands.

[0050] Based on the updated mapping parameters The system converts high-frequency data from the angular motion sensor into continuous pointing coordinates, and encapsulates operations such as pointing coordinates, clicking, long pressing, dragging, scrolling, or returning into standard input commands for output to the display terminal, host device, or application. Preferably, the output method can be Bluetooth HID, USB HID, proprietary wireless protocol, or application layer communication protocol.

[0051] (iv) Preferred hardware implementation method In a preferred embodiment, the handheld pointing device includes a main control chip, a dual-imaging-channel polarization image acquisition module, an angular velocity sensor, a wireless communication module, a button or touch input module, and a power management module. The field of view of the two imaging channels covers the display area at common user operating distances, and polarizers with different polarization analysis directions are respectively set in front of the two channels.

[0052] A gyroscope is preferred as the angular motion sensor. An accelerometer is not essential for pointing coordinate calculations but can be used as an optional module for pick-up wake-up, stationary detection, power consumption control, and coarse attitude assessment, avoiding cursor positioning issues caused by acceleration integral drift. The visual observation frequency can be lower than the angular motion data sampling frequency, with the specific frequency dynamically adjusted based on power consumption, processing power, display area, user operation speed, and visual confidence level.

[0053] (iv) Preferred hardware implementation method In a preferred embodiment, the handheld pointing device includes a main control chip, a dual-imaging-channel polarization image acquisition module, an angular velocity sensor, a wireless communication module, a button or touch input module, and a power management module. The field of view of the two imaging channels covers the display area at common user operating distances, and polarizers with different polarization analysis directions are respectively set in front of the two channels.

[0054] A gyroscope is preferred as the angular motion sensor. An accelerometer is not essential for pointing coordinate calculations but can be used as an optional module for pick-up wake-up, stationary detection, power consumption control, and coarse attitude assessment, avoiding cursor positioning issues caused by acceleration integral drift. The visual observation frequency can be lower than the angular motion data sampling frequency, with the specific frequency dynamically adjusted based on power consumption, processing power, display area, user operation speed, and visual confidence level.

Claims

1. A pointing interaction method based on polarization visual anchoring, characterized in that, Includes the following steps: S1. Acquire image data of the displayed scene under at least two different polarization analysis states using a handheld pointing device; S2. Based on the image data from the at least two different polarization analysis states, calculate the polarization response characteristics used to distinguish the display area from the environmental background; S3. Perform dynamic threshold segmentation based on the polarization response feature map, and filter out candidate display areas in combination with geometric constraints; S4. Establish a mapping relationship between the image coordinate system and the display area coordinate system based on the selected display area to obtain the visual pointing coordinates; S5. Collect angular motion data at a sampling frequency higher than the visual observation frequency, and convert the angular motion change into the display area coordinate change between adjacent visual observation cycles according to the current mapping parameters, and output high-frequency predicted pointing coordinates. S6. When the change in visual pointing coordinates obtained at two consecutive visual observation times meets the preset confidence condition, the mapping parameter between the change in visual coordinates and the integral of angular motion within the corresponding time period is dynamically updated using the change in visual coordinates and the integral of angular motion within the corresponding time period. S7. Based on the updated mapping parameters, convert the angular motion data into continuous pointing coordinates, and encapsulate them into interactive commands for output to the display terminal.

2. The method according to claim 1, characterized in that, The polarization response feature is at least one of the following: normalized differential feature, brightness difference, brightness ratio, polarization degree approximation, or response consistency feature under multiple polarization states; when the normalized differential feature is used, its calculation formula is: ; in , These are the brightness values ​​under two different polarization analysis states. To prevent positive numbers with a denominator of zero.

3. The method according to claim 1, characterized in that, The geometric constraints include one or more combinations of the following: connected region area range, boundary continuity, quadrilateral shape fitting error of candidate regions, aspect ratio range, number of corner points, boundary straightness, and temporal consistency with the display area position of the previous frame; the determination of the display area candidates adopts a combined judgment method of "polarization response difference + geometric constraints + temporal confidence".

4. The method according to claim 1, characterized in that, Dynamically updating mapping parameters uses a recursive filtering method: ; in Change in current visual coordinates Integral quantity of angular motion in the same time period Calculations show that The updated weights are adaptively adjusted based on visual confidence, angular motion amplitude, or angular velocity noise level, and ; Wherein, the angular motion integral quantity The mapping parameters can be a vector containing changes in yaw and pitch angles. It can be a scalar scaling factor, a two-dimensional matrix, a local linear mapping matrix, or a nonlinear correction parameter; When using a two-dimensional angular motion integral vector It can be obtained by using least squares, recursive least squares, or local fitting with regularization terms from the changes in visual coordinates and integrals of angular motion over multiple consecutive visual observation cycles.

5. The method according to claim 1, characterized in that, It also includes confidence assessment and anomaly handling steps: A comprehensive confidence score is calculated for each visual observation result. The comprehensive confidence score is determined by at least two of the following factors: polarization response difference intensity, candidate region area, boundary straightness, corner stability, aspect ratio rationality, and consistency with historical position. When the overall confidence level is higher than the first threshold, the visual observation results are used to update the current pointing coordinates and mapping parameters; When the overall confidence level is lower than the second threshold, the mapping parameter update is paused, and only the previous effective mapping parameter is used for short-term angular motion prediction. Boundary constraints or velocity damping are applied to the predicted coordinates until the visual observation returns to stability. When the overall confidence level is between the first and second thresholds, reduce the weight of visual observations on the update of mapping parameters, or use only visual observations for slow correction.

6. A pointing interaction device based on polarization visual anchoring, characterized in that, include: The polarization image acquisition module is used to acquire image data of the display scene under at least two different polarization analysis states; The display area recognition module is used to extract the display area based on polarization response differences and geometric constraints. The coordinate mapping module is used to establish a screen coordinate system or a pointing coordinate mapping relationship based on the observation results of the display area. An angular motion acquisition module is used to acquire the angular velocity or angular motion change of a handheld device. The dynamic calibration module is used to update the mapping parameters from angular motion to screen coordinate displacement based on continuous visual observation and synchronous angular motion integral. The interactive output module is used to output cursor coordinates, click, drag, or scroll interaction commands.

7. The apparatus according to claim 6, characterized in that, The polarization image acquisition module has any of the following structures: Two imaging channels with fixed polarization analysis directions, the fields of view of the two channels are basically overlapping, and polarization analyzers with different polarization directions are set respectively; An integrated polarization image sensor acquires pixel responses from multiple different polarization analysis directions on the same imaging chip; A single imaging channel, in conjunction with an electronically controlled polarization switching element, can be switched to at least two different polarization analysis states in time division, and an angular motion sensor is used to compensate for attitude changes between time division images.

8. The apparatus according to claim 6, characterized in that, The angular motion acquisition module is a gyroscope, and the device does not use the integral result of the accelerometer as a necessary input for the calculation of the pointing coordinates; the accelerometer is only an optional module used for device wake-up, static state detection, power consumption control or coarse attitude judgment.

9. The apparatus according to claim 6, characterized in that, It also includes a power consumption control module: When the pointing function is not enabled, the device is stationary, or there is no effective angular movement for a long time, the sampling frequency of the polarization image acquisition module is reduced or it is put into a sleep state. When the pointing function is detected to be enabled, the device is picked up, or the angular motion amplitude exceeds the preset threshold, the image acquisition frequency is restored or increased. During the pointing function, if the cursor does not move effectively or the visual confidence level is higher than the preset threshold, the visual observation frequency is dynamically reduced; if the angular motion amplitude increases, the visual confidence level decreases, or the cursor approaches the boundary of the display area, the visual observation frequency is temporarily increased.

10. The apparatus according to claim 6, characterized in that, The instruction format output by the interactive output module is Bluetooth HID instruction, USB HID instruction, private wireless protocol instruction, or application layer communication protocol instruction; the device is a handheld pointing device, and its form includes, but is not limited to, a TV remote control, a conference screen pointing device, an educational whiteboard controller, a demonstration controller, or a smart home pointing controller.

Images