Multi-point touch screen touch recognition method and system based on capacitive coupling optimization

Abstract

The application relates to the technical field of capacitive coupling, in particular to a touch screen multi-point touch recognition method and system based on capacitive coupling optimization, which comprises the following steps: collecting electrode array signals, decomposing and calculating errors, screening interference signals, generating a recognition result, dividing a grid based on the recognition result, normalizing and removing signals to generate a distribution graph, collecting glass back signals, sorting, layering and marking to generate a hierarchical graph, extracting intensity changes, calculating directions, screening areas, aggregating to generate a grouping graph, obtaining projection coordinates, combining levels and aggregating to calculate the gravity center to generate a structure graph. In the application, the data purity is improved by screening interference signals through frequency domain decomposition, false signals are removed through medium layer grid division and normalization, the touch point area is enhanced through charge response layering and marking, the boundary is ensured to be clear through intensity gradient direction aggregation, the positioning accuracy is improved through space projection and gravity center calculation, and the recognition accuracy and anti-interference ability are improved as a whole.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of capacitive coupling, in particular to a multi-point touch recognition method and system for a touch screen based on capacitive coupling optimization. BACKGROUND

[0002] The technical field of capacitive coupling includes touch detection, electric signal collection and position determination, and its core content is to establish an electric field coupling relationship through capacitive coupling to sense and recognize the approach or contact of an input object to the screen. It generally involves the detection of capacitive changes, the optimization of coupling paths, and the analysis of touch signals, covering the structural design of capacitive sensors, charge transmission methods, electrode layout schemes, and recognition mechanisms for coupled signals, and is an important technical basis for human-computer interaction of touch screens.

[0003] Among them, the multi-point touch recognition method and system for a touch screen based on capacitive coupling optimization refers to establishing multiple capacitive coupling channels in the sensing electrodes of the touch screen, and distinguishing and collecting the coupling signals of multiple touch points in combination with the capacitive change characteristics. In view of the signal interference and coupling ambiguity under multi-point touch, the technical matters covered include the partition design of capacitive coupling paths, the synchronous collection of capacitive signals, and the cross-determination of touch positions, and specifically, the distribution optimization of electrode matrices and the capacitive channel coupling correction method are used for multi-point touch recognition.

[0004] In the prior art, the recognition of multi-point touch mainly relies on the distribution optimization of electrode matrices and the coupling correction of capacitive channels, but due to the complex interference in the transmission process of signals, the original signals are difficult to keep pure, resulting in ambiguity or even misjudgment of touch points at the boundary position. The detection mode of capacitive change is prone to signal overlap when multiple touch points exist at the same time, making the separation between multiple touch points not clear enough. The transmission of charge signals in the dielectric layer is limited by the spatial distribution law, and the existing methods lack deeper analysis of charge gradient and spatial hierarchy, resulting in insufficient precision of touch point positioning. In the multi-point touch situation, two fingers close to each other are often recognized as a single touch point, or the touch trajectory drifts when moving quickly, which directly affects the experience of the touch screen in high-precision applications such as games and drawing, and reduces the response sensitivity and interaction reliability of the system. SUMMARY

[0005] To solve the technical problems existing in the prior art, the embodiments of the present application provide a multi-point touch recognition method for a touch screen based on capacitive coupling optimization, which comprises the following steps:

[0006] In order to achieve the above purpose, the present application adopts the following technical scheme: a multi-point touch recognition method for a touch screen based on capacitive coupling optimization, comprising the following steps:

[0007] S1: Collect the original charge signal of the transparent electrode array and perform frequency domain decomposition, calculate the charge response error of adjacent charge signals by least square method and fitting, screen the charge signal pairs exceeding the coupling interference threshold, and generate a coupling interference identification result;

[0008] S2: Based on the coupling interference identification result, the touch screen medium layer is obtained and divided into a grid, the charge intensity change of adjacent sampling points is calculated and normalized to remove the charge signal caused by capacitive coupling, and a decoupling charge distribution map is generated;

[0009] S3: Collect the charge response corresponding to the decoupled charge distribution map on the back of the glass and sort it, divide the region with monotonically decreasing charge response amplitude into multiple layers and add layer labels, and generate a touch point space level label map;

[0010] S4: Based on the touch point space level label map, the charge intensity change is extracted, and the vector direction of the charge intensity gradient is calculated, the touch point area exceeding the gradient amplitude threshold is screened for grouping and aggregation, and a touch point direction grouping map is generated;

[0011] S5: Based on the touch point direction grouping map, the spatial projection coordinates of the multi-group area are obtained, the touch control barycenter position is calculated by combining the touch point space level label map and the corresponding touch point aggregation area, and a multi-point touch control recognition structure map is generated.

[0012] As a further scheme of the application, the coupling interference identification result includes coupling offset, interference signal pair, the decoupled charge distribution map is specifically charge intensity distribution, normalization coefficient, the touch point space level label map includes level sequence number, region boundary coordinates, layer label set, the touch point direction grouping map specifically refers to vector direction set, gradient amplitude interval, aggregation grouping index, and the multi-point touch control recognition structure map includes barycenter coordinate set, spatial projection, grouping mapping relationship.

[0013] As a further scheme of the application, the specific steps of S1 are:

[0014] S101: Collect the original charge signal of the transparent electrode array, perform frequency domain decomposition on the collected time series signal, calculate the frequency component value after frequency domain decomposition and classify, and generate a frequency component value;

[0015] S102: Based on the frequency component value, the adjacent charge signals are fitted by least square method, the charge response error between the fitted curve and the current charge signal is calculated and summarized as a sequence, and a charge response error sequence is generated;

[0016] S103: According to the charge response error sequence, call the coupling interference threshold to screen the multiple error coefficients one by one, extract the signal pairs exceeding the coupling interference threshold and collect them, and generate a coupling interference identification result.

[0017] As a further scheme of the present application, the coupling interference threshold is set by taking the energy distribution value of the extracted charge signal in the multi-frequency interval, and taking the average and standard deviation of the energy distribution value as the basis, and setting according to the average plus a multiple of the standard deviation greater than 1.

[0018] As a further scheme of the present application, the specific steps of S2 are:

[0019] S201: Based on the coupling interference identification result, the area of the touch screen medium layer is obtained, the touch screen medium layer area is grid divided, and the charge intensity change data of adjacent sampling points is collected;

[0020] S202: Based on the charge intensity change data, the charge intensity change between each adjacent sampling point is normalized, the charge signal caused by capacitive coupling is eliminated, and a decoupled charge signal is generated;

[0021] S203: According to the decoupled charge signal, the signal distribution after elimination is mapped to the medium layer grid to generate a decoupled charge distribution map.

[0022] As a further scheme of the present application, the specific steps of S3 are:

[0023] S301: Collecting charge response data corresponding to the decoupled charge distribution map on the back of the glass, sorting all charge response amplitudes according to amplitude size to generate a charge response amplitude sequence;

[0024] S302: Based on the charge response amplitude sequence, detecting the range of the monotonically decreasing amplitude interval, segmenting the interval according to the amplitude change trend, and assigning a hierarchical label to the multi-segment to generate multi-layer division annotation data;

[0025] S303: According to the multi-layer division annotation data, the multi-segment hierarchical information is correspondingly mapped to the spatial position of the back of the glass to generate a touch point spatial hierarchical annotation map.

[0026] As a further scheme of the present application, the specific steps of S4 are:

[0027] S401: Based on the touch point spatial hierarchical annotation map, the charge intensity change of each touch point area is extracted, the gradient of the charge intensity change is calculated, the gradient amplitude and direction of each touch point area are analyzed and combined to obtain a charge intensity gradient vector;

[0028] S402: According to the charge intensity gradient vector, the gradient amplitude of each touch point area is screened to exclude touch point areas that do not reach a set gradient amplitude threshold to generate screened touch point areas;

[0029] S403: Group the vector directions of each contact area based on the screened contact areas, aggregate the contact areas with the same direction, and generate a contact direction grouping map.

[0030] As a further scheme of the present application, the gradient amplitude threshold is set by calculating the overall distribution interval of the gradient amplitude based on the statistical analysis of the contact area charge intensity variation range, and then setting according to the median and standard deviation of the distribution interval.

[0031] As a further scheme of the present application, the specific steps of S5 are:

[0032] S501: Based on the contact direction grouping map, obtain the boundary point coordinates of the multi-grouping area, expand and map the boundary point coordinates according to the grouping direction, calculate the projection position of each grouping area in two-dimensional space, and generate a space projection coordinate;

[0033] S502: Call the space projection coordinate, combine the contact space level label map recorded in the contact level position, weight the projection coordinates of the multi-contact area and the level coordinates, calculate the center point position of each contact aggregation area as the touch gravity center coordinates;

[0034] S503: According to the touch gravity center coordinates and the corresponding contact aggregation area, spatially map and integrate the gravity center coordinates of the multi-grouping area to generate a multi-point touch recognition structure map.

[0035] The multi-point touch recognition system of the capacitive coupling optimized touch screen comprises:

[0036] The charge analysis module collects the original charge signal of the transparent electrode array and performs frequency domain decomposition, calculates and fits the charge response error of adjacent charge signals by least square method, screens the charge signal pairs exceeding the coupling interference threshold, generates a coupling interference recognition result and transmits it to the charge decoupling module;

[0037] The charge decoupling module obtains the touch screen medium layer based on the coupling interference recognition result and divides it into a grid, calculates the charge intensity variation of adjacent sampling points, and normalizes and removes the charge signal caused by capacitive coupling to generate a decoupled charge distribution map and transmit it to the contact layering module;

[0038] The contact layering module collects and sorts the charge response corresponding to the decoupled charge distribution map on the back of the glass, divides the region with monotonically decreasing charge response amplitude into multiple layers and adds a layering label, generates a contact space level label map and transmits it to the direction aggregation module;

[0039] A direction aggregation module extracts charge intensity variation based on the contact spatial hierarchical labeling map, calculates vector directions of charge intensity gradient, screens contact regions exceeding a gradient amplitude threshold for grouping aggregation, generates a contact direction grouping map, and passes the contact direction grouping map to a gravity center recognition module;

[0040] The gravity center recognition module obtains spatial projection coordinates of multi-group regions based on the contact direction grouping map, calculates a touch control gravity center position in combination with the contact spatial hierarchical labeling map and corresponding contact aggregated regions, and generates a multi-point touch control recognition structure map.

[0041] Compared with the prior art, the application has the following advantages and positive effects:

[0042] In the application, interference signal pairs are screened through frequency domain decomposition and error fitting of the least square method, accurate identification of abnormal charge coupling is performed, interference signals are avoided from being misjudged in subsequent processing, and the purity of original data is improved. Further, after the medium layer is divided into grids, the change of charge intensity of adjacent sampling points is normalized, false signals caused by coupling effects can be eliminated, and the touch control map is clearer and more distinguishable. On this basis, the charge response collected on the back of the glass is labeled in layers, so that the touch points have clear separation in the spatial hierarchy, and the resolution capability of multiple touch points under multi-point touch control is enhanced. Through vector analysis and aggregation grouping of the charge intensity gradient direction, the touch points are reasonably divided in terms of directionality and regionality, and the signal boundary is still clear under complex touch control. Finally, the spatial hierarchical labeling and the coordinates of the aggregated regions are combined to calculate the touch control gravity center, so that the determination of the multi-point touch control position is more accurate and stable. The overall processing logic under the continuous actions of gradually eliminating interference, enhancing separation, optimizing grouping and accurate positioning makes the conversion of touch control recognition from original signal to spatial coordinates more complete and rigorous, significantly improves the recognition accuracy and anti-interference ability of multi-point touch control, and improves the sensitivity and reliability of touch control feedback in complex environments. BRIEF DESCRIPTION OF DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0044] Figure 1 The step flowchart of the present application is shown in the figure;

[0045] Figure 2 The S1 refinement schematic diagram of the present application is shown in the figure;

[0046] Figure 3 The S2 refinement schematic diagram of the present application is shown in the figure;

[0047] Figure 4 This is a detailed schematic diagram of S3 of the present invention;

[0048] Figure 5 This is a detailed schematic diagram of S4 of the present invention;

[0049] Figure 6 This is a detailed schematic diagram of S5 of the present invention;

[0050] Figure 7 This is a system module diagram of the present invention. Detailed Implementation

[0051] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0052] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0053] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0054] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0055] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0056] Please see Figure 1 This invention provides a multi-touch recognition method for touchscreens based on capacitive coupling optimization, comprising the following steps:

[0057] S1: Acquire the original charge signal of the transparent electrode array and perform frequency domain decomposition. Calculate the charge response error of adjacent charge signals using the least squares method and fit the result. Filter out charge signal pairs that exceed the coupling interference threshold and generate coupling interference identification results.

[0058] S2: Based on the coupling interference identification results, obtain the touch screen medium layer and divide it into grids, calculate the charge intensity change of adjacent sampling points and normalize it to remove the charge signal caused by capacitive coupling, and generate a decoupled charge distribution map;

[0059] S3: Collect and sort the charge response corresponding to the decoupling charge distribution map on the back of the glass, divide the areas with monotonically decreasing charge response amplitude into multiple layers and attach layer labels to generate a contact space hierarchy labeling map.

[0060] S4: Extract the charge intensity change based on the spatial hierarchy map of the contact points, calculate the vector direction of the charge intensity gradient, filter the contact point regions that exceed the gradient amplitude threshold, group and aggregate them, and generate a contact point direction group map;

[0061] S5: Obtain the spatial projection coordinates of multiple grouped regions based on the touch point direction grouping map, calculate the touch center position by combining the touch point spatial hierarchy annotation map and the corresponding touch point aggregation area, and generate a multi-point touch recognition structure map.

[0062] The coupling interference identification results include coupling offset and interference signal pairs. The decoupling charge distribution map specifically includes charge intensity distribution and normalization coefficient. The touch point spatial hierarchy labeling map includes hierarchy sequence number, region boundary coordinates, and layer label set. The touch point direction grouping map specifically refers to vector direction set, gradient amplitude range, and aggregation grouping index. The multi-touch recognition structure map includes centroid coordinate set, spatial projection, and grouping mapping relationship.

[0063] Please see Figure 2 The specific steps of S1 are as follows:

[0064] S101: Acquire the original charge signal of the transparent electrode array, perform frequency domain decomposition on the acquired time series signal, calculate and classify the frequency component values ​​after frequency domain decomposition, and generate frequency component values.

[0065] At the third row and fourth column of an 8x8 transparent electrode array, a raw charge signal was acquired for one second at a sampling rate of 5000 Hz, resulting in a time series of 5000 data points. The acquired signal sequence was {0.21, 0.24, 0.29, 0.35, 0.31} picocoulms. Subsequently, this time series was decomposed in the frequency domain. This process did not directly call a pre-defined algorithm library, but instead calculated the component values ​​at specific frequency points one by one. Taking the calculation of the amplitude of the 20 Hz frequency component as an example, first, each data point in the time series was multiplied by the corresponding 20 Hz cosine waveform value and summed to obtain the accumulated real part value; then, each data point was multiplied by the 20 Hz sine waveform value and summed to obtain the accumulated imaginary part value. The square root of the sum of the squares of the accumulated real and imaginary parts was then taken to obtain the amplitude of the 20 Hz frequency component. The calculated amplitudes were 9.2 picocoulombs at 10 Hz, 7.5 picocoulombs at 50 Hz, and 2.1 picocoulombs at 150 Hz. After calculating all frequency points from 1 Hz to 2500 Hz, the obtained frequency component values ​​were categorized. The classification criteria were preset to four frequency bands: low-frequency activity band (1-12 Hz), mid-frequency signal band (12-100 Hz), high-frequency noise band (100-500 Hz), and ultra-high-frequency interference band (500-2500 Hz). The total energy of the calculated frequency component amplitudes within each band was obtained by summing the squares. The generated frequency component values ​​were a set, specifically: total energy of the low-frequency activity band 256.3 (picocoulombs)², total energy of the mid-frequency signal band 4102.1 (picocoulombs)², total energy of the high-frequency noise band 315.8 (picocoulombs)², and total energy of the ultra-high-frequency interference band 98.4 (picocoulombs)².

[0066] S102: Based on the frequency component values, the adjacent charge signals are fitted using the least squares method, the charge response error between the fitted curve and the current charge signal is calculated and summarized into a sequence, generating a charge response error sequence;

[0067] Based on the obtained total energy of 256.3 (picocoulon)² for the low-frequency activity band, 4102.1 (picocoulon)² for the mid-frequency signal band, 315.8 (picocoulon)² for the high-frequency noise band, and 98.4 (picocoulon)² for the ultra-high-frequency interference band, the original charge signals of the electrodes in the 3rd row and 4th column were obtained, and the charge signals of the adjacent electrodes in the 3rd row and 5th column were simultaneously acquired. These two signals were fitted using the least squares method. A linear relationship was established: the fitted signal value of the electrode in the 3rd row and 5th column equals a gain coefficient multiplied by the signal value of the electrode in the 3rd row and 4th column, plus a bias constant. To determine the gain coefficient and bias constant, the sum of the squares of the differences between the current signal value and the fitted signal value of the electrode in the 3rd row and 5th column at all sampling time points needs to be minimized. The calculations are as follows: the sum of the pointwise products of the two electrode signal sequences, the sum of the squares of the multiples of the two signals, and the sum of all multiples of the data points. Taking a signal with 5 points as an example, let the electrode signal in the 3rd row and 4th column be {0.2, 0.4, 0.6, 0.5, 0.3} picocoulons, and the electrode signal in the 3rd row and 5th column be {0.14, 0.25, 0.35, 0.30, 0.20} picocoulons. Through the above summation operation, the gain coefficient is calculated to be 0.52, and the bias constant is 0.04. This yields the fitting relationship. Next, the charge response error between the fitted curve and the current charge signal is calculated. That is, at each sampling time point, the current signal value of the electrode in the 3rd row and 5th column is subtracted from the fitted signal value calculated by the relationship. For example, at the first time point, the error is 0.14 - (0.52 * 0.2 + 0.04), resulting in -0.004 picocoulons. The error values ​​calculated at all time points are summarized in chronological order to form a data sequence containing 5000 error values; this sequence is the charge response error sequence.

[0068] S103: Based on the charge response error sequence, the coupling interference threshold is called to filter the multiple error coefficients one by one, extract the signal pairs that exceed the coupling interference threshold and collect them to generate coupling interference identification results;

[0069] Based on the charge response error sequence with 5000 error values ​​generated in the preceding steps, a preset coupling interference threshold is used for signal pair screening. The threshold setting process is as follows: On a blank electrode array without biological samples, a simulated signal with known parameters is injected into one electrode using a microprobe, and the response signals of its adjacent electrodes are recorded. This process is repeated 200 times, with the amplitude and frequency of the injected signal varying randomly within a preset range each time. For each recorded response signal and injected signal, least squares fitting and error sequence calculation are performed in S102, and the root mean square value of the error sequence is calculated. These 200 root mean square values ​​are statistically analyzed, and the arithmetic mean is calculated to be 0.25 picocoulombs, with a standard deviation of 0.08 picocoulombs. The coupling interference threshold is set to the mean plus three times the standard deviation, i.e., 0.25 plus 3 multiplied by 0.08, which is determined to be 0.49 picocoulombs. The threshold setting ensures the statistical significance of the judgment. Subsequently, the root mean square (RMS) value of the charge response error sequence for the current measurement sample (electrode pair in row 3, column 4 and column 3, column 5) is calculated. Each value in the error sequence is squared, the average of all squared values ​​is calculated, and then the square root of the average is taken. The calculated RMS value of the error sequence is set to 0.61 picocoulombs. This value is compared with the coupling interference threshold: 0.61 picocoulombs is greater than 0.49 picocoulombs. Based on this, the signal pair (electrode pair in row 3, column 4 and column 3, column 5) is determined to exceed the coupling interference threshold. The identification information of all signal pairs exceeding the threshold, such as their coordinate pairs in the array, is aggregated to generate (E4, 5; E4, 6) and (E2, 1; E2, 2).

[0070] Please see Figure 3 The specific steps of S2 are as follows:

[0071] S201: Based on the coupling interference identification results, obtain the region of the touch screen medium layer, divide the touch screen medium layer region into grids, and collect the charge intensity change data of adjacent sampling points;

[0072] Based on the coupling interference identification results generated in the preceding steps, namely the lists (E4, 5; E4, 6) and (E2, 1; E2, 2), the relevant areas of the touch screen's dielectric layer are first located and divided. The coupling interference identification results indicate the specific locations of significant signal crosstalk in the electrode array. These electrode coordinates are mapped onto the dielectric layer of a capacitive touch screen with physical dimensions of 150 mm x 80 mm. The electrode pair (E4, 5; E4, 6) corresponds to an area centered at a point (x=62, y=55) mm in the screen coordinate system. After obtaining this central area, a 20 mm x 20 mm square area is defined as the focus of processing. A refined virtual mesh is performed on this dielectric layer area, setting the unit size of the mesh to 0.5 mm x 0.5 mm, thus dividing the 20 mm x 20 mm area into a 40 x 40 mesh. Each intersection of this mesh is defined as a sampling point, totaling 1600 sampling points. Subsequently, in the initial non-touch state, a charge intensity scan was performed on all sampling points within the area, recording the reference charge value for each point. The reference charge intensity for sampling point P(20, 22) was set to 1.25 picocoulombs (pC), and the reference charge intensity for the adjacent sampling point P(20, 23) was set to 1.28 picocoulombs. Next, a standardized metal stylus with a diameter of 5 mm was used to contact the screen with a constant pressure of 1.5 Newtons, with the center of the contact point located near sampling point P(20, 22). At the instant of stylus contact, a charge intensity scan was performed again on all sampling points within the area to obtain the charge intensity values ​​under touch conditions. The charge intensity at sampling point P(20, 22) under touch conditions became 1.85 picocoulombs, and at sampling point P(20, 23) it became 1.62 picocoulombs. By subtracting the reference charge value of the corresponding point from the charge intensity value under touch conditions, the charge intensity change data for each sampling point was calculated. For sampling point P(20, 22), the charge intensity change is 1.85 - 1.25 = 0.60 pC. For the adjacent sampling point P(20, 23), the charge intensity change is 1.62 - 1.28 = 0.34 pC. This acquisition process is repeated for all 1600 sampling points and all adjacent sampling point pairs (horizontally and vertically) within this 40x40 grid area to obtain a dataset of charge intensity changes between adjacent sampling points covering the entire area.

[0073] S202: Based on the charge intensity change data, the charge intensity change between each adjacent sampling point is normalized to remove the charge signal caused by capacitive coupling and generate a decoupled charge signal.

[0074] After acquiring the charge intensity variation data covering the entire 40x40 grid area, normalization and decoupling operations are performed on the charge intensity variation of each sampling point and its adjacent sampling points. First, the entire dataset is traversed to find the maximum and minimum values ​​of charge intensity variation among all 1600 sampling points. The maximum charge intensity variation value is set to 0.65 pC (sampling point directly below the stylus), and the minimum value is set to 0.01 pC (edge ​​sampling point far from the touch point). The normalization process is as follows: for any sampling point, the normalized charge intensity variation value is equal to (the original charge intensity variation value at that point minus the minimum value of 0.01 pC), then divided by (the maximum value of 0.65 pC minus the minimum value of 0.01 pC). Taking sampling point P(20, 22) as an example, the change in original charge intensity is 0.60 pC, and the normalized value is (0.60-0.01) / (0.65-0.01)=0.59 / 0.64≈0.922. For its adjacent sampling point P(20, 23), the change in original charge intensity is 0.34 pC, and the normalized value is (0.34-0.01) / (0.65-0.01)=0.33 / 0.64≈0.516. Next, the normalized signal is decoupled based on the coupling gain coefficient calculated for the regional electrode pair in S102. In the previous steps, the gain coefficient α of the linear fitting model for the electrode pair (E4, 5; E4, 6) was calculated to be 0.47. This coefficient characterizes the strength of signal coupling between adjacent electrodes (i.e., adjacent sampling points in this case). The specific action to remove the charge signal caused by capacitive coupling is as follows: for a pair of adjacent sampling points A and B, the decoupled signal value at point B is equal to the original signal value at point B minus the product of the original signal value at point A and the coupling gain coefficient 0.47. This operation is based on the premise that the point with higher signal strength (point A) will have a major coupling effect on the adjacent point with lower signal strength (point B). Here, the signal strength of P(20, 22) is higher than that of P(20, 23), so P(20, 22) is regarded as the coupling source and P(20, 23) is regarded as the coupled point. The decoupled charge signal (normalized value) at P(20, 23) is calculated as follows: 0.516 - (0.922 × 0.47) = 0.516 - 0.433 = 0.083. This decoupling calculation is applied to all adjacent sampling point pairs within the region that have a primary-secondary strength relationship, generating a set of decoupled charge signals covering the entire region.

[0075] S203: Based on the decoupled charge signal, the signal distribution after removal is mapped onto the dielectric layer mesh to generate a decoupled charge distribution map;

[0076] Based on generating a set of decoupled charge signals covering the entire region, and simultaneously acquiring decoupled charge signals covering the entire 40x40 grid area, the signal distribution after removing coupling effects is remapped onto the virtual grid of the dielectric layer. The mapping process is a point-by-point assignment operation. A two-dimensional data matrix is ​​created that perfectly corresponds to the 40x40 grid divided by the dielectric layer. The row and column indices of the matrix directly correspond to the grid coordinates. The decoupled charge signal value calculated for each sampling point is filled into the corresponding coordinate position in this two-dimensional data matrix. For example, the coordinates of sampling point P(20, 22) correspond to the 20th row and 22nd column of the matrix, and the value assigned to this position is its normalized signal value of 0.922. The adjacent sampling point P(20, 23) corresponds to the 20th row and 23rd column of the matrix, and the value assigned is the decoupled signal value of 0.083. The assignment operation is performed on all 1600 sampling points until the entire 40x40 two-dimensional data matrix is ​​filled.

[0077] Table 1 Comparison of Signal Values ​​in Local Grid Regions

[0078]

[0079] As shown in Table 1, this table lists the normalized signal values ​​of the contact center P(20, 22) and some of its adjacent sampling points before and after decoupling. The original signal values ​​show that the signal strength smoothly decays outward from the contact center, forming a blurry "bright spot". The decoupled signal values ​​show that the signal strength at the center point remains unchanged, but the signal strength at the adjacent points is significantly weakened. This result indicates that the higher signal values ​​measured at the adjacent points were mostly transmitted from the strong signal at the center point through capacitive coupling, rather than caused by actual touch pressure. Visualizing the filled two-dimensional data matrix generates a decoupled charge distribution map. The distribution map before processing appears as a large, blurry spot, while the generated decoupled charge distribution map appears as a spot with a highly prominent center point, clear edges, and a significantly reduced area.

[0080] Please see Figure 4 The specific steps of S3 are as follows:

[0081] S301: Collect charge response data corresponding to the decoupled charge distribution map on the back of the glass, sort all charge response amplitudes according to their magnitude, and generate a charge response amplitude sequence;

[0082] First, charge response data corresponding to the 40x40 decoupled charge distribution map generated in the previous steps is collected from the back of the glass. This acquisition process directly reads all 1600 values ​​in this distribution map (a 40x40 two-dimensional data matrix). These values ​​represent the normalized charge intensity of each grid sampling point after the touch event, excluding capacitive coupling effects. For subsequent processing, this data needs to be converted from a two-dimensional spatial structure into a one-dimensional sequence. First, a validity threshold for the charge response is set, based on statistical analysis of the background noise level collected by the system in a touchless state. By measuring 1000 frames of data in a touchless state, the mean of the noise signal amplitude at all sampling points is calculated to be 0.0005, and the standard deviation is 0.0008. The validity threshold is set to the mean plus three times the standard deviation, i.e., 0.0005 + 3 × 0.0008 = 0.0029. The 1600 decoupled charge signal values ​​in the 40x40 matrix are traversed, filtering out all values ​​greater than 0.0029. After filtering, 52 valid charge response data points are obtained. The charge response amplitude values ​​of these 52 data points are extracted, forming an unordered set. For example, some values ​​in this set might be {0.922, 0.083, 0.101, 0.062, 0.045, 0.075, ...}. While extracting the values, the two-dimensional coordinates (x, y) of each value in the original 40x40 grid are recorded. Next, these 52 charge response amplitudes are sorted in descending order of magnitude. This sorting process does not call specific library functions but is performed as follows: First, the maximum value is found among the 52 values ​​and used as the first element of the sequence; then, the maximum value is found among the remaining 51 values ​​and used as the second element of the sequence; this process is repeated until all values ​​are placed into the new sequence. After sorting, a charge response amplitude sequence is generated. For example, the first 10 elements of the sequence are: {0.922, 0.101, 0.083, 0.081, 0.075, 0.062, 0.058, 0.045, 0.044, 0.041}. Each element in this sequence is implicitly associated with its original spatial coordinates.

[0083] S302: Based on the charge response amplitude sequence, detect the range of intervals with monotonically decreasing amplitude, divide the intervals into segments according to the amplitude change trend, and assign hierarchical labels to the multiple segments to generate multi-level segmentation and labeling data;

[0084] Based on the generated charge response amplitude sequence {0.922, 0.101, 0.083, 0.081, 0.075, 0.062, 0.058, 0.045, 0.044, 0.041}, the range of monotonically decreasing amplitude intervals is detected, and the intervals are divided according to the amplitude change trend. Since the sequence itself is already in descending order, the entire sequence is a large monotonically decreasing interval. The key to this division is identifying the "inflection point" where the decreasing trend changes significantly. Specifically, the difference between every two adjacent elements in the sequence is calculated to obtain an "amplitude decay" sequence. For example, the amplitude decay values ​​for the first few elements of the sequence are calculated as follows: the first decay value is 0.922 - 0.101 = 0.821; the second is 0.101 - 0.083 = 0.018; the third is 0.083 - 0.081 = 0.002; and the fourth is 0.081 - 0.075 = 0.006. Subsequent calculations result in an amplitude decay sequence of 51 elements: {0.821, 0.018, 0.002, 0.006, ...}. Next, one or more "attenuation abruptness thresholds" need to be set to segment the interval. The threshold was set based on experimental calibration of the amplitude attenuation sequence generated when multiple touch objects (such as conductive rubber tips, fingertips, and metal needle tips) touched the screen under standard pressure.

[0085] Table 2 Experimental data on amplitude attenuation characteristics of multiple touch objects

[0086]

[0087] As shown in Table 2, experimental data shows that the amplitude attenuation value from the core contact point to the transition area (e.g., 0.750) differs by orders of magnitude from the amplitude attenuation value from the transition area to the edge area (e.g., 0.045) and the attenuation value within the same area (e.g., 0.004). Based on this experimental data, two segmentation thresholds were set: a first-level segmentation threshold of 0.200 and a second-level segmentation threshold of 0.025. The amplitude attenuation sequence was traversed, and the values ​​were compared with the thresholds. In the attenuation sequence {0.821, 0.018, 0.002, ...}, the first value 0.821 is much larger than the first-level segmentation threshold of 0.200, indicating a hierarchical breakpoint between the first and second elements of the sequence. In subsequent calculations of the sequence, the 6th attenuation value (i.e., the difference between the 7th and 8th elements of the original sequence) was set to 0.028, which is less than 0.200 but greater than the second-level segmentation threshold of 0.025, indicating another hierarchical breakpoint at this point. Based on these two breakpoints, the original charge response amplitude sequence is divided into three segments. The first segment includes only the first element {0.922}. The second segment includes elements from the second to the seventh {0.101, 0.083, 0.081, 0.075, 0.062, 0.058}. The third segment includes all remaining elements starting from the eighth element. These three segments are assigned hierarchical labels: the first segment is labeled "Level 1: Core Contact"; the second segment is labeled "Level 2: Main Deformation Zone"; and the third segment is labeled "Level 3: Edge Sensing Zone", generating multi-level segmentation and labeling data.

[0088] S303: Based on the multi-layer division and annotation data, the multi-segment hierarchical information is mapped to the spatial position on the back of the glass to generate a touch point spatial hierarchy annotation map;

[0089] Based on the generated multi-layered labeled data, the first segment is labeled "Level 1: Core Contact Point"; the second segment is labeled "Level 2: Main Deformation Zone"; and the third segment is labeled "Level 3: Edge Sensing Zone". The hierarchical information of these segments is mapped to their original spatial positions on the back of the glass. This process requires utilizing the original two-dimensional coordinates stored in S301 for each charge response amplitude value. First, a 40x40 two-dimensional empty matrix with the same size as the touchscreen medium layer grid is created, with all elements initialized to 0, representing "non-responsive areas". Then, each data segment with a hierarchical label is processed sequentially. For the segment labeled "Level 1: Core Contact Point", which includes an amplitude value of 0.922, the original coordinates recorded in S301 are found to be (20, 22). Therefore, in the 40x40 empty matrix, the value of the element in row 20, column 22 is changed from 0 to 1. Next, the segment labeled "Level 2: Main Deformation Zone" is processed. This segment includes six amplitude values: {0.101, 0.083, 0.081, 0.075, 0.062, 0.058}. The original coordinates corresponding to these six values ​​are queried one by one and set to (21, 22), (20, 23), (19, 22), (20, 21), (22, 22), and (18, 22). In the 40x40 matrix, the element values ​​at these six coordinate positions are changed from 0 to 2. Finally, the segment labeled "Level 3: Edge Sensing Zone" is processed. This segment includes the remaining 45 amplitude values. Similarly, the original coordinates of each of these 45 values ​​are queried, and the element values ​​at the corresponding positions in the matrix are changed from 0 to 3. Once the coordinate lookup and matrix assignment operations for all amplitude values ​​with level labels have been completed, the filling of the entire 40x40 matrix is ​​finished. It reprojects the one-dimensional signal strength hierarchical analysis results back into a two-dimensional physical space, directly constructing a structured spatial representation of the touch point area by assigning each spatial point a discrete hierarchical identity rather than a continuous intensity value. This 40x40 two-dimensional matrix filled with hierarchical labels (0, 1, 2, 3) is the generated touch point spatial hierarchy annotation map.

[0090] Please see Figure 5 The specific steps of S4 are as follows:

[0091] S401: Extract the charge intensity change of each contact area based on the contact space hierarchy annotation map, calculate the gradient of charge intensity change, analyze the gradient magnitude and direction of each contact area and combine them to obtain the charge intensity gradient vector;

[0092] Based on the generated contact spatial hierarchy map, all contact regions labeled as level 1, 2, or 3 are extracted, and the charge intensity variation values ​​of multiple sampling points within these regions are retrieved from the decoupled charge distribution map. The charge intensity gradient is calculated for each sampling point within each extracted contact region. Taking sampling point P(20, 21) in the "level 2" region as an example, its decoupled charge intensity value is 0.062. The gradient calculation requires the charge intensity values ​​of adjacent points: right P(21, 21) is 0.050, left P(19, 21) is 0.055, above P(20, 22) is 0.922, and below P(20, 20) is 0.010. The physical size of the grid cell is 0.5 mm. The horizontal gradient component of this point is equal to the difference between the charge intensity of the right point and the charge intensity of the left point, divided by the physical distance of 1.0 mm between the two points, calculated as follows: Units: mm. The vertical gradient component is equal to the difference between the intensity of the upper point charge and the intensity of the lower point charge, divided by the physical distance between the two points, 1.0 mm, and calculated as follows: Units: mm. The gradient magnitude is calculated as the square root of the sum of the squares of the horizontal and vertical gradient components, i.e. Units: mm. The gradient direction is the arctangent of the vertical and horizontal gradient components, calculated as atan2(vertical gradient component, horizontal gradient component) = atan2((0.922-0.010) / 1.0, (0.050-0.055) / 1.0) = atan2(0.912, -0.005), resulting in 90.3°. Combining this gradient magnitude and direction yields the charge intensity gradient vector at point P(20, 21). This process is repeated for all sampled points within the extracted contact area, generating a set consisting of multi-point coordinates and corresponding charge intensity gradient vectors.

[0093] S402: Based on the charge intensity gradient vector, filter the gradient amplitude of each contact area, exclude contact areas that do not reach the set gradient amplitude threshold, and generate the filtered contact areas.

[0094] Based on the generated set of charge intensity gradient vectors, the charge intensity gradient vector at point P(20, 21) is obtained. A filtering operation is then performed on the gradient amplitude for each contact area. A gradient amplitude threshold is established experimentally. In the experiment, 200 measurements were performed on hovering gestures without physical contact. The recorded noise gradient amplitude had a mean of 0.048 units / mm and a standard deviation of 0.021 units / mm. The gradient amplitude threshold was set to the noise mean plus four times the standard deviation, i.e. The unit is per millimeter. Based on this, the gradient magnitude is divided into two intervals: less than 0.132 is the "low gradient noise region," and greater than or equal to 0.132 is the "high gradient effective region." The filtering process is as follows: traverse each gradient vector in the set, extract its magnitude, and compare it with the threshold 0.132. For example, the gradient magnitude of point P(20, 21) calculated in S401 is 0.912 units / mm. Since 0.912 is greater than 0.132, this point belongs to the "high gradient effective region," and the contact area is retained. The gradient magnitude of edge point P(28, 30) is calculated to be 0.110 units / mm. Since 0.110 is less than 0.132, this point belongs to the "low gradient noise region," and the contact area is excluded from the set. This filtering judgment is applied to all gradient vectors in the set, generating a new set that only includes contact areas with gradient magnitudes reaching the set threshold; this is the filtered contact area.

[0095] S403: Based on the filtered contact areas, group the vector directions of each contact area, aggregate contact areas with the same direction, and generate a contact direction grouping diagram;

[0096] Based on the generated set of filtered touchpoint regions, the gradient vector directions of each touchpoint region are grouped and aggregated. First, the 360-degree directional range is divided into eight fixed directional intervals, each 45 degrees wide. Specifically, the intervals are: Group 1 (East) [-22.5°, 22.5°], Group 2 (Northeast) 22.5°, 67.5°, Group 3 (North) 67.5°, 112.5°, Group 4 (Northwest) 112.5°, 157.5°, Group 5 (West) 157.5°, 202.5°, Group 6 (Southwest) 202.5°, 247.5°, Group 7 (South) 247.5°, 292.5°, and Group 8 (Southeast) 292.5°, 337.5°. The grouping process involves iterating through each gradient vector in the filtered set, extracting its direction angle, and determining which directional interval the angle falls into. Taking the gradient vector of point P(20, 21) as an example, its direction angle is 90.3°, which falls within the range of 67.5° to 112.5°. Therefore, this contact area is classified into "Group Three (North)". Another contact area in the set, P(19, 22), has a gradient vector direction angle of 95.1°, which also falls within this range and is classified into "Group Three (North)". After all contact areas are classified by direction, contact areas belonging to the same direction group are aggregated. {Group Three (North): [(20, 21), (19, 22), ...], Group One (East): [(21, 23), ...], generating a contact direction grouping diagram.}

[0097] Please see Figure 6 The specific steps of S5 are as follows:

[0098] S501: Obtain the boundary point coordinates of multiple grouped regions based on the contact direction grouping map, expand and map the boundary point coordinates according to the grouping direction, calculate the projection position of each grouped region in two-dimensional space, and generate spatial projection coordinates;

[0099] Based on the generated contact point direction grouping map, i.e., the data structure {Group 3 (North): [(20, 21), (19, 22), ...], Group 1 (East): [(21, 23), ...], the coordinates of the boundary points are first obtained for each group region. Taking "Group 3 (North)" as an example, the set of coordinate points included in this group is {(20, 21), (19, 22), (18, 22), (20, 23)}. The criterion for extracting boundary points is: the point with the largest coordinate value in the "North" direction (90°) defined by the group. In this set, the point with the largest y-coordinate is (20, 23), therefore, (20, 23) is determined as the boundary point of "Group 3 (North)". The coordinates of this boundary point are then expanded and mapped according to its group direction. This mapping process involves a preset "expansion coefficient", which is used to determine the projection distance. The expansion factor was determined through the following experiment: Multiple circular conductive contacts of various sizes (3 mm to 10 mm in diameter) were used to contact the screen with a constant pressure of 1.5 Newtons. The number of sampling points in the contact aggregation area generated by each contact was recorded, and the current physical radius of the contact was measured. By fitting 500 sets of data on the "number of sampling points" and "physical radius," a relationship was determined: the physical radius is approximately equal to 0.8 mm multiplied by the square root of the number of sampling points. Based on this, the expansion factor was set to 0.8. For the "Group 3 (North)" region, which includes 4 sampling points, considering that the physical size of the grid cell is 0.5 mm, the physical radius relationship was converted to grid units: grid radius. The grid unit is used, therefore the distance of the unfolded mapping is calculated as follows: Each grid unit is grouped in the direction of "North," i.e., the positive y-axis, with a direction vector of (0, 1). The specific calculation for the unfolding mapping is as follows: the boundary point coordinates (20, 23) are added to the product of the direction vector (0, 1) and the unfolding distance 1.6. The calculation process is as follows: the new x-coordinates... New y-coordinate The projected position of the grouped region in two-dimensional space is calculated, and its spatial projection coordinates are (20, 24.6). For all grouped regions in the contact direction grouping diagram (e.g., "Group 1 (East)", "Group 5 (West)", etc.), the boundary point extraction, unfolding distance calculation and unfolding mapping operations are repeated to generate a set consisting of the spatial projection coordinates of multiple grouped regions.

[0100] S502: Call the spatial projection coordinates, combine them with the touch level positions recorded in the touch spatial hierarchy annotation diagram, weight the projection coordinates of the multi-touch area with the hierarchy coordinates, calculate the center point position of each touch aggregation area, and use it as the touch center coordinates;

[0101] The generated set of spatial projection coordinates is used, combined with the multi-touch level position information recorded in the generated touch point spatial hierarchy annotation map, to perform weighted calculations on the coordinates of the multi-touch area. The goal is to calculate the center point position of each complete touch point aggregation area. First, weight values ​​need to be assigned to multiple touch point levels. These weight values ​​reflect the importance of each level in determining the touch center of gravity. The weight settings are derived from calibration experiments on touch pressure distribution.

[0102] Table 3. Experimental Data for Contact Level Weight Calibration

[0103]

[0104] As shown in Table 3, the physical pressure corresponding to multiple layers of regions is simultaneously measured by a micro-pressure sensor array when a touch occurs. The average pressure of the "Layer 1" region is normalized to 1.00 to obtain the relative pressure values ​​of other layers. To enhance the contribution of the core region and widen the gap between layers, the normalized pressure value is multiplied by an amplification factor of 3 to obtain the final set weights: Layer 1 weight is 3.0, Layer 2 weight is 1.5, and Layer 3 weight is 0.5. A weighted average is calculated for the coordinates of all sampling points in a touch aggregation area (i.e., all 48 valid touch points selected by S402). The advantage of this method is that by introducing layer weights directly related to physical pressure, the calculated center of gravity can more accurately reflect the focus of the user's intent, rather than a simple geometric center. Taking a micro-aggregated area comprising three points as an example, this area includes: core touch point P1 (20, 22), with a level of 1 and a weight of 3.0; main deformation area touch point P2 (21, 22), with a level of 2 and a weight of 1.5; and edge sensing area touch point P3 (21, 21), with a level of 3 and a weight of 0.5. The calculation process for the touch center coordinates is as follows: First, calculate the sum of the weights of all points: Then, calculate the weighted sum of the x-coordinates: Next, calculate the weighted sum of the y-coordinates: Finally, calculate the x-coordinate of the centroid: Calculate the y-coordinate of the centroid: Applying this calculation process to all 48 touch point areas in this touch event yields the center point of the entire touch point aggregation area, i.e., the touch centroid coordinates.

[0105] S503: Based on the touch center coordinates and the corresponding touch aggregation area, the center coordinates of the multi-grouped areas are spatially mapped and integrated to generate a multi-touch recognition structure diagram;

[0106] Based on the calculated touch centroid coordinates and the corresponding entire touch aggregation area, the centroid information of the multi-grouped areas is spatially mapped and integrated. Multiple independent touch events occurring simultaneously are processed, and a structured final output is generated. First, connected component analysis is performed on all activated touch points on the screen to distinguish multiple independent touch aggregation areas. When two touches are detected simultaneously on the screen, connected component analysis identifies 48 points filtered by S402 as belonging to the first touch aggregation area, and another 35 points as belonging to the second independent touch aggregation area. Calculation S502 is performed on the first area, yielding touch centroid coordinates of (20.4, 21.9). The complete calculation process from S501 to S502 is performed independently on the second area, setting its touch centroid coordinates to (35.8, 40.2). Next, these two independent centroid coordinates are spatially mapped and integrated. The integration process creates a structured data record containing complete information for each identified independent touch event. Each touch event record includes the following fields: a unique touch event ID; calculated touch centroid coordinates; the set of coordinates of all sampling points constituting the touch event; and boundary information of the touch area. The boundary information can be defined by the spatial projection coordinates calculated in S501 for multiple directions. For example, these projection points can form a convex polygon surrounding the touch area. All this information is combined to generate a multi-touch recognition structure diagram. Below is example data for a structure diagram including two simultaneous touch events: Touch Event 1: Touch ID: 001, centroid coordinates: (20.4, 21.9), touch point set: {(20, 22), (21, 22), (21, 21), ..., a total of 48 points}, boundary projection points: {(20, 24.6), (23.5, 22), (18, 19.4), ...}, Touch Event 2: Touch ID: 002, centroid coordinates: The set of touch points (35.8, 40.2) is {(35, 40), (36, 40), (35, 41), ..., a total of 35 points}, and the boundary projection points are {(35, 42.1), (38.2, 40), (33.9, 39), ...}. This result shows that the system simultaneously recognized two independent touch operations and accurately located their weighted centers, while also providing the complete set of points constituting each touch and its spatial contour information. This structural diagram provides comprehensive and accurate input data for subsequent gesture recognition or user interface interaction.

[0107] Please see Figure 7 A multi-touch recognition system for touchscreens based on capacitive coupling optimization includes:

[0108] The charge analysis module acquires the original charge signal of the transparent electrode array and performs frequency domain decomposition. It calculates and fits the charge response error of adjacent charge signals using the least squares method, filters charge signal pairs that exceed the coupling interference threshold, generates coupling interference identification results, and transmits them to the charge decoupling module.

[0109] The charge decoupling module obtains the touch screen medium layer based on the coupling interference identification results and divides it into a grid. It calculates the charge intensity change of adjacent sampling points and normalizes it to remove the charge signal caused by capacitive coupling. It generates a decoupled charge distribution map and transmits it to the touch point layering module.

[0110] The contact layering module collects and sorts the charge response corresponding to the decoupling charge distribution map on the back of the glass. It divides the areas with monotonically decreasing charge response amplitude into multiple layers and adds layer labels to generate a contact spatial hierarchy annotation map and transmits it to the direction aggregation module.

[0111] The orientation aggregation module extracts the charge intensity change based on the spatial hierarchy annotation map of the contact points, calculates the vector direction of the charge intensity gradient, filters the contact point regions that exceed the gradient amplitude threshold, groups and aggregates them, generates a contact point orientation group map, and transmits it to the centroid recognition module.

[0112] The center of gravity recognition module obtains the spatial projection coordinates of multiple grouped areas based on the touch point direction grouping map, and calculates the touch center of gravity position by combining the touch point spatial hierarchy annotation map and the corresponding touch point aggregation area, thereby generating a multi-point touch recognition structure map.

[0113] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A multi-touch recognition method for touchscreens based on capacitive coupling optimization, characterized in that, Includes the following steps: S1: Acquire the original charge signal of the transparent electrode array and perform frequency domain decomposition. Calculate the charge response error of adjacent charge signals using the least squares method and fit the result. Filter out charge signal pairs that exceed the coupling interference threshold and generate coupling interference identification results. S2: Based on the coupling interference identification results, obtain the touch screen medium layer and divide it into grids, calculate the charge intensity change of adjacent sampling points and normalize it to remove the charge signal caused by capacitive coupling, and generate a decoupled charge distribution map; S3: Collect the charge response on the back of the glass corresponding to the decoupled charge distribution map and sort it. Divide the regions with monotonically decreasing charge response amplitudes into multiple layers and add layer labels to generate a contact space hierarchy labeling map. S4: Extract the charge intensity change based on the contact spatial hierarchy map, calculate the vector direction of the charge intensity gradient, filter the contact areas that exceed the gradient amplitude threshold, group and aggregate them, and generate a contact direction grouping map; S5: Based on the touch point direction grouping map, obtain the spatial projection coordinates of the multi-grouped areas, combine the touch point spatial hierarchy annotation map with the corresponding touch point aggregation area to calculate the touch center position, and generate a multi-touch recognition structure map; The specific steps of S5 are as follows: S501: Based on the contact point direction grouping diagram, obtain the boundary point coordinates of the multi-grouped regions, expand and map the boundary point coordinates according to the grouping direction, calculate the projection position of each grouped region in two-dimensional space, and generate spatial projection coordinates; S502: Call the spatial projection coordinates, combine them with the touch point level positions recorded in the touch point spatial level annotation diagram, weight the projection coordinates of the multi-touch point area with the level coordinates, calculate the center point position of each touch point aggregation area, and use it as the touch center coordinates; S503: Based on the touch center coordinates and the corresponding touch aggregation area, the center coordinates of the multi-grouped areas are spatially mapped and integrated to generate a multi-touch recognition structure diagram.

2. The multi-touch recognition method for touch screens based on capacitive coupling optimization according to claim 1, characterized in that, The coupling interference identification result includes coupling offset and interference signal pair; the decoupling charge distribution map specifically includes charge intensity distribution and normalization coefficient; the touch point spatial hierarchy labeling map includes hierarchy sequence number, region boundary coordinates, and layer label set; the touch point direction grouping map specifically refers to vector direction set, gradient amplitude interval, and aggregation grouping index; and the multi-touch recognition structure map includes centroid coordinate set, spatial projection, and grouping mapping relationship.

3. The multi-touch recognition method for touch screens based on capacitive coupling optimization according to claim 1, characterized in that, The specific steps of S1 are as follows: S101: Acquire the original charge signal of the transparent electrode array, perform frequency domain decomposition on the acquired time series signal, calculate and classify the frequency component values ​​after frequency domain decomposition, and generate frequency component values. S102: Based on the frequency component values, the adjacent charge signals are fitted using the least squares method, the charge response error between the fitted curve and the current charge signal is calculated and summarized into a sequence to generate a charge response error sequence; S103: Based on the charge response error sequence, the coupling interference threshold is called to filter the multiple error coefficients one by one, extract the signal pairs that exceed the coupling interference threshold and collect them to generate the coupling interference identification result.

4. The multi-touch recognition method for touch screens based on capacitive coupling optimization according to claim 3, characterized in that, The coupling interference threshold is set by extracting the energy distribution value of the charge signal in multiple frequency ranges, and then using the average and standard deviation of the energy distribution value as a benchmark, based on the average plus a standard deviation greater than 1.

5. The multi-touch recognition method for touch screens based on capacitive coupling optimization according to claim 1, characterized in that, The specific steps of S2 are as follows: S201: Based on the coupling interference identification result, obtain the region of the touch screen medium layer, divide the touch screen medium layer region into a grid, and collect the charge intensity change data of adjacent sampling points; S202: Based on the charge intensity change data, normalize the charge intensity change between each adjacent sampling point, remove the charge signal caused by capacitive coupling, and generate a decoupled charge signal; S203: Based on the decoupled charge signal, the signal distribution after removal is mapped onto the dielectric layer grid to generate a decoupled charge distribution map.

6. The multi-touch recognition method for touch screens based on capacitive coupling optimization according to claim 1, characterized in that, The specific steps for S3 are as follows: S301: Collect charge response data on the back of the glass corresponding to the decoupled charge distribution map, sort all charge response amplitudes according to their magnitude, and generate a charge response amplitude sequence; S302: Based on the charge response amplitude sequence, detect the range of intervals where the amplitude decreases monotonically, divide the intervals into segments according to the amplitude change trend, and assign hierarchical labels to the multiple segments to generate multi-level segmentation and labeling data; S303: Based on the multi-layer division and annotation data, map the multi-segment hierarchical information with the spatial position on the back of the glass to generate a touch point spatial hierarchy annotation map.

7. The multi-touch recognition method for touch screens based on capacitive coupling optimization according to claim 1, characterized in that, The specific steps of S4 are as follows: S401: Based on the contact point spatial hierarchy annotation map, extract the charge intensity change of each contact point region, calculate the gradient of charge intensity change, analyze the gradient magnitude and direction of each contact point region and combine them to obtain the charge intensity gradient vector; S402: Based on the charge intensity gradient vector, filter the gradient amplitude of each contact area, exclude contact areas that do not reach the set gradient amplitude threshold, and generate the filtered contact areas. S403: Based on the filtered contact areas, group the vector directions of each contact area, aggregate contact areas with the same direction, and generate a contact direction grouping diagram.

8. The multi-touch recognition method for touch screens based on capacitive coupling optimization according to claim 7, characterized in that, The gradient amplitude threshold is determined by statistically analyzing the range of charge intensity variation in the contact area, calculating the overall distribution range of the gradient amplitude, and then setting it based on the median and standard deviation of the distribution range.

9. A multi-touch recognition system for touchscreens based on capacitive coupling optimization, characterized in that, The system is used to perform the multi-touch recognition method for a touch screen based on capacitive coupling optimization as described in any one of claims 1-8, and the system includes: The charge analysis module acquires the original charge signal of the transparent electrode array and performs frequency domain decomposition. It calculates and fits the charge response error of adjacent charge signals using the least squares method, filters charge signal pairs that exceed the coupling interference threshold, generates coupling interference identification results, and transmits them to the charge decoupling module. The charge decoupling module obtains the touch screen medium layer and divides it into grids based on the coupling interference identification results, calculates the charge intensity change of adjacent sampling points and normalizes it to remove the charge signal caused by capacitive coupling, generates a decoupled charge distribution map and transmits it to the touch point layering module. The contact layering module collects and sorts the charge response on the back of the glass corresponding to the decoupled charge distribution map, divides the areas with monotonically decreasing charge response amplitude into multiple layers and adds layering labels, generates a contact spatial hierarchy labeling map and transmits it to the direction aggregation module. The orientation aggregation module extracts the charge intensity change based on the contact spatial hierarchy annotation map, calculates the vector direction of the charge intensity gradient, filters contact areas that exceed the gradient amplitude threshold for group aggregation, generates a contact orientation group map, and transmits it to the centroid recognition module. The center of gravity recognition module obtains the spatial projection coordinates of multiple grouped regions based on the touch point direction grouping map, calculates the touch center of gravity position by combining the touch point spatial hierarchy annotation map and the corresponding touch point aggregation area, and generates a multi-point touch recognition structure map.

Images