JP6218228B2 - Touch detection device and touch detection method - Google Patents

Description

  The present invention relates to a touch detection device and a touch detection method that can detect a change in data associated with a touch more accurately and improve the accuracy of touch coordinates based on low data input from the touch panel side.

  In recent years, a projection capacitive method has been widely used for touch panels in information terminals from the viewpoint of excellent transmissivity and durability in correspondence with multi-touch (for example, see Non-Patent Document 1). ). FIG. 3 is an explanatory diagram of a projection-type capacitance method that is a conventional technique. In this projection type capacitive system, a drive electrode (Tx) and a reception electrode (Rx), which are sensor electrodes, are arranged orthogonally on the upper surface of the display, and a sensor is formed at each intersection (see FIG. 3A). Then, a change in capacitance between the two electrodes that occurs when the panel surface is touched with a finger or the like is detected from a difference in sensor output signals with and without touch (see FIGS. 3B and 3C).

  FIG. 4 is an overall configuration diagram including a conventional touch detection device, and includes a touch panel 110, a preprocessing unit 120, and a touch detection device 130. An analog signal that is a sensor output signal from the touch panel 110 is taken into the preprocessing unit 120. Then, the preprocessing unit 120 performs necessary analog signal processing and AD conversion to generate raw data (raw data) that is two-dimensional digital data for each scan period.

  Then, the subsequent touch detection device 130 generates difference data in which noise and offset are reduced and removed from the raw data. Then, using the difference data generated in this way, the presence / absence of touch is detected, the coordinates of the touch position are calculated, and the like.

  As the noise filter, there are a recursive filter corresponding to random noise for each two-dimensional element, and an offset removing filter corresponding to random noise for each row (Tx) when a higher effect is expected. Used.

  Here, the difference data is obtained by subtracting baseline data from the low data. In the conventional touch detection device, this baseline data is a signal after noise reduction. Then, the conventional touch detection device sets the initial row data as the initial state and increases or decreases the value of each element of the data for each frame so that the baseline data asymptotically approaches the row data in the state where the value does not change due to the touch. Processing such as updating was performed.

Overview of flat panel display (6) Parts and material technology supporting FPD (1) Touch panel, THE CHEMICAL TIMES 2011, No. 4

However, the prior art has the following problems.
Recently, there has been a strong demand for touch input with a pen as well as a finger. However, in the projection-type capacitance method, when a thin conductive touch instrument such as a pen is used, the change in capacitance in the sensor, that is, the change in sensor output is the case of a finger. For example, it is very weak with about 1/10.

  Therefore, with the noise reduction measures in the conventional touch detection device corresponding to the touch with the finger, in the case of the touch with the pen, the noise becomes relatively large, and the erroneous detection or the deviation of the calculated coordinate is very It becomes easy to happen.

  Further, as an essential problem, there is a kind of noise that is not taken into consideration in the noise reduction measures of the conventional touch detection device. That is, fixed pattern noise (FPN: Fixed-Pattern Noise) superimposed on data in a fixed and steady manner with variations in characteristics for each sensor position and for each Tx circuit or Rx circuit has not been considered.

  The present invention has been made in order to solve the above-described problems, and it is possible to appropriately reduce and remove superposition noise and offset with respect to the low data input from the touch panel side, thereby changing data accompanying touch. It is an object of the present invention to obtain a touch detection device and a touch detection method that can detect more accurately and improve the calculation accuracy of touch coordinates.

  A touch detection device according to the present invention is a touch detection device that outputs a touched position coordinate on a panel based on low data, which is sensor output data after AD conversion from a touch panel, for each sensor position and sensor output. Fixed pattern noise filter that outputs intermediate difference data obtained by subtracting background data from low data after removing fixed pattern noise that is superimposed on sensor output data in a fixed and steady manner due to variations in characteristics of each data generation circuit A post-stage filter that outputs differential data by reducing random noise depending on the sensor position included in the sensor output data, and a difference output from the post-stage filter, with respect to the intermediate differential data output from the fixed pattern noise filter Touch determination unit for outputting touch determination data based on the data, and subsequent filter And a coordinate calculation unit that outputs the touched position coordinates based on the difference data output from the touch determination unit and the touch determination data output from the touch determination unit, and the fixed pattern noise filter includes a first control command from the outside In accordance with the second control command from the outside, a weighted average is obtained by varying the weight for averaging the frame data for storing the latest background data, the latest background data read from the frame memory, and the raw data. A weighted averaging unit that outputs the converted background data, a subtractor that outputs intermediate difference data by subtracting the weighted averaged background data output from the variable weighted averaging unit from the raw data, and touch determination Based on the touch determination data output from the first section, the background data after weighted averaging is stored in the frame memory as the latest background data. It generates a control command, and is intended to be configured to include a filter control unit that generates a second control command for obtaining the background data after weighted averaging with the desired weight.

  In addition, the touch detection method according to the present invention can be performed in a fixed and steady manner from the low data, which is sensor output data after AD conversion from the touch panel, with variations in characteristics for each sensor position and for each sensor output data generation circuit. Included in sensor output data for fixed pattern noise filter that outputs intermediate difference data obtained by subtracting background data from low data and intermediate difference data output from fixed pattern noise filter after removing fixed pattern noise to be superimposed A post-filter that outputs differential data by reducing random noise depending on the sensor position, a touch determination unit that outputs touch determination data based on the differential data output from the post-filter, and a difference output from the post-filter Based on the data and touch determination data output from the touch determination unit A touch detection method that is applied to a touch detection device that outputs a touched position coordinate on a panel, the coordinate calculation unit outputting the touched position coordinate. In response to the first control command, an update step for updating the latest background data stored in the frame memory, a background data extraction step for reading the latest background data stored in the frame memory, and the latest read out from the frame memory. The weight for averaging the background data and the raw data is varied according to the second control command from the outside, and is output in the variable weighted averaging step and the variable weighted averaging step for outputting the background data after weighted averaging. Subtracting step to output intermediate difference data by subtracting the background data after weighted averaging from low data, and touch From the touch determination unit, a first generation step for generating a first control command for storing the background data after weighted averaging in the frame memory as the latest background data based on the touch determination data output from the fixing unit A second generation step of generating a second control command for obtaining background data after weighted averaging with a desired weight based on the output touch determination data.

  According to the present invention, by providing a filter that reduces and eliminates fixed pattern noise that is superimposed on data in a fixed and steady manner with variations in characteristics for each sensor position and for each Tx circuit or Rx circuit, Touch detection device and touch detection that can detect the change of data accompanying touch more accurately and improve the calculation accuracy of touch coordinates by appropriately reducing and removing superposition noise and offset for the input low data You can get the method.

It is a block diagram of the touch detection apparatus in Embodiment 1 of this invention. In Embodiment 1 of this invention, it is the figure which showed the structural example of the touch detection apparatus different from FIG. It is explanatory drawing of the projection-type electrostatic capacitance system which is a prior art. It is a whole block diagram containing the conventional touch detection apparatus.

  Hereinafter, preferred embodiments of a touch detection device and a touch detection method of the present invention will be described with reference to the drawings.

Embodiment 1 FIG.
FIG. 1 is a configuration diagram of a touch detection device according to Embodiment 1 of the present invention. The touch detection apparatus according to the first embodiment shown in FIG. 1 includes a fixed pattern noise filter 10, a line-by-line noise filter 20, a recursive noise filter 30, a touch determination unit 40, and a coordinate calculation unit 50. Yes.

  And the touch detection apparatus in this Embodiment 1 has such a structure, and calculates the coordinate data of the touched position with high accuracy by using raw data from a pre-processing unit (not shown) as an input. Can do. Therefore, each component will be described in detail below.

  The fixed pattern noise filter 10 includes a variable recursive filter 11 including a variable weighted averaging unit 11a and a frame memory 11b, a variable recursive filter control unit 12, and a subtractor 13. The background data is generated from the raw data by the variable recursive filter 11, the updated background data is stored in the frame memory 11b, and the subtractor 13 subtracts the background data from the raw data to generate intermediate difference data.

  That is, the fixed pattern noise filter 10 extracts intermediate difference data for detecting that there is a touch by comparing background data in a state where there is no touch with low data that is sequentially input.

  The variable recursive filter control unit 12 generates a signal for controlling the variable recursive filter 11 based on the touch determination data from the touch determination unit 40. Then, the variable recursive filter control unit 12 changes the weighting of the raw data in the variable weighted averaging unit 11a and the latest background data extracted from the frame memory 11b by the generated control signal (control command), and The valid / stop of data writing / updating in the frame memory 11b is switched.

  The operation of the fixed pattern noise filter 10 having such a configuration for removing fixed pattern noise will be described more specifically. First, immediately after initialization, the variable recursive filter control unit 12 sets the background data to be the same as the raw data by setting the weight of the average value calculation of the low data and the background data in the variable weighted averaging unit 11a to 1: 0. The frame memory 11b can be written, and background data (= low data) is written into the frame memory 11b. By performing such an operation, the raw low data not touched is stored in the frame memory 11b as background data.

  The touch determination unit 40 passes through the row-by-row noise filter 20 and the recursive noise filter 30, so that there is a touch based on the value of each element of the difference data after the offset removal and noise reduction. It is determined whether or not.

  Then, the variable recursive filter control unit 12 counts the number of consecutive non-touch frames in which the number of elements determined to be touched is zero, and controls the variable recursive filter 11 based on the count value. Do.

  Specifically, the variable recursive filter control unit 12 determines that when the number of consecutive frames without touch is less than a predetermined number (for example, 5 frames) (this includes when there is a touched frame, that is, when there is no touch (Including the case where the number of consecutive frames is 0), the weight of the average value calculation of the raw data and the background data in the variable weighted averaging unit 11a is set to 0: 1 and the update of the background data in the frame memory 11b is stopped. To do.

  As a result, the variable recursive filter 11 prohibits updating the data in the frame memory 11b with the low data after the touch has occurred, and the frame without touch immediately before the touch determination is stored, which is stored in the frame memory 11b. Continue to output the background data. As a result, the subtractor 13 can generate intermediate difference data based on the touched low data and the background data of the non-touch frame.

  The low data and the background data in this state both include fixed pattern noise that is superimposed on the data in a fixed and steady manner with variations in characteristics for each sensor position and for each Tx circuit or Rx circuit. . Therefore, the influence of fixed pattern noise can be removed by taking the difference between the two data.

  On the other hand, when the number of consecutive frames without touch reaches a predetermined number, the variable recursive filter control unit 12 sets the weight of the average value calculation of the raw data and background data in the variable weighted averaging unit 11a to 1: 0. Thus, the raw data is used as the background data as it is, and the update of the background data in the frame memory 11b is validated.

  Further, when the number of consecutive touchless frames is greater than a predetermined number, the variable recursive filter control unit 12 sets the weight for calculating the average value of the raw data and the background data in the variable weighted averaging unit 11a, for example, 1: 1. Alternatively, background data is generated as 1: 3, and the frame memory 11b is updated with this data.

  At this time, as the background data is updated, random noise is reduced in the latest background data stored in the frame memory 11b by the variable weighted averaging process, and fixed pattern noise and offset remain mainly. . Therefore, by subtracting the background data generated in this way from the raw data and canceling out noise components that are commonly included in the two data to generate intermediate difference data, the fixed pattern noise and the offset are reduced or eliminated.

  As a result, in the case of a frame with a touch, random noise for each row, each column, and each element of frame data is mainly superimposed on the value increased with the touch. On the other hand, in the case of a non-touch frame, random noise is the main component. Then, differential data from which random noise has been removed is generated by a post-stage filter including the row-by-line noise filter 20 and the recursive noise filter 30 provided at the rear stage of the fixed pattern noise filter 10.

  More specifically, the intermediate difference data is then weighted in FIG. 1 by a row averaging unit 21, a line memory 22, a noise filter 20 per row having a subtractor 23, and a frame memory 31. Through the recursive noise filter 30 having the averaging unit 32, random noise for each row and random noise for each data element are reduced. Here, the row averaging unit 21 excludes elements determined to be touched from the average value calculation.

  Thus, based on the difference data after reducing and removing various noises and offsets, the touch determination unit 40 determines the presence or absence of a touch, and the coordinate calculation unit 50 calculates touch coordinates. Further, the coordinate calculation unit 50 obtains a tracking ID considering the temporal continuity of the touch coordinates, and outputs the coordinates together with the ID.

  In addition to the recursive noise filter 30, a two-dimensional low-pass filter (two-dimensional LPF) can be applied to reduce random noise for each element of the frame data. Further, when the random noise for each column is significant, the random noise for each column can be reduced by using a column-by-column noise filter having a column averaging unit and a frame memory.

  FIG. 2 is a diagram showing a configuration example of a touch detection device different from that in FIG. 1 in Embodiment 1 of the present invention. That is, the row-by-row noise filter 20 is used to remove random noise for each row, but the row-by-column noise filter 20a can be used to remove random noise for each column. 20 and the noise filter 20a for each column can be used in combination.

  In order to remove random noise for each element of the frame data, the recursive noise filter 30 or the two-dimensional LPF 30a can be used. An example of these combinations is shown in FIGS. 2A to 2C, and a post-stage filter for removing random noise is provided in various stages after the fixed pattern noise filter 10. realizable.

  As described above, according to the first embodiment, in the touch detection device that outputs the coordinates of the touched position on the panel based on the raw data input from the touch panel through the preprocessing unit, the fixed pattern noise and the offset It is equipped with a fixed pattern noise filter that can reduce and eliminate noise.

  More specifically, the fixed pattern noise filter includes a variable recursive filter having a variable weighted averaging unit and a frame memory, and a filter control unit for controlling the variable recursive filter based on continuous touch determination results. And a subtractor. By providing such a configuration, it is possible to generate background data and to generate intermediate difference data in which background data is subtracted from low data to reduce and remove fixed pattern noise and offset.

  Further, a noise filter capable of generating difference data by reducing random noise for each row, each column, and each sensor position is provided at the subsequent stage of the fixed pattern noise filter that generates intermediate difference data. As a result, it is possible to realize a touch detection device capable of determining the presence / absence of touch and calculating the touched coordinates with higher accuracy based on the difference data obtained by appropriately reducing and removing noise and offset.

  10 fixed pattern noise filter, 11 variable recursive filter, 11a variable weighted averaging unit, 11b frame memory, 12 variable recursive filter control unit, 13 subtractor, 20 row noise filter, 20a column noise filter, 21 row average Conversion unit, 22 line memory, 23 subtractor, 30 recursive noise filter, 30a two-dimensional LPF, 31 frame memory, 32 weighted averaging unit, 40 touch determination unit, 50 coordinate calculation unit.

Claims (4)

A touch detection device for outputting a touched position coordinate on a panel based on low data which is sensor output data after AD conversion from a touch panel,
Intermediate pattern obtained by subtracting background data from the low data after removing fixed pattern noise that is superimposed on the sensor output data in a fixed and steady manner due to variations in the characteristics of each sensor position and sensor output data generation circuit. A fixed pattern noise filter that outputs differential data;
For the intermediate difference data output from the fixed pattern noise filter, a post-filter that reduces the random noise depending on the sensor position included in the sensor output data and outputs the difference data;
A touch determination unit that outputs touch determination data based on the difference data output from the subsequent filter;
A coordinate calculation unit that outputs the touched position coordinates based on the difference data output from the subsequent filter and the touch determination data output from the touch determination unit;
The fixed pattern noise filter is
A frame memory for storing the latest background data in response to a first control command from the outside;
Variable weighted averaging for varying the weight for averaging the latest background data read from the frame memory and the raw data in accordance with an external second control command and outputting the background data after weighted averaging And
A subtractor that outputs the intermediate difference data by subtracting the background data after the weighted averaging output from the variable weighted averaging unit from the raw data;
Based on the touch determination data output from the touch determination unit, generates the first control command for storing the background data after the weighted average in the frame memory as the latest background data, and desired And a filter control unit that generates the second control command for obtaining the background data after the weighted averaging by the weight of the touch detection device. The touch detection device according to claim 1,
The filter control unit
In the touch determination data, count the number of consecutive non-touch frames in which the number of elements determined to be touched is zero,
When the number of consecutive non-touch frames is less than a predetermined number, the second control command is generated so that the weight of the average value calculation of the low data and the background data is 0: 1, and the frame Generating the first control command to stop the update of the background data in the memory;
When the number of consecutive frames without touch reaches the predetermined number, the second control command is generated so that the weight of the average value calculation is 1: 0, and the background data in the frame memory is updated. The touch detection device that generates the first control command so as to be effective. The touch detection device according to claim 1, wherein:
The post-stage filter includes at least one of a row-by-row noise filter, a column-by-column noise filter, a recursive noise filter, and a two-dimensional low-pass filter. After removing fixed pattern noise that is superimposed on the sensor output data from the touch panel, which is sensor output data after A / D conversion, in a fixed and steady state with variations in characteristics for each sensor position and each sensor output data generation circuit. A fixed pattern noise filter that outputs intermediate difference data obtained by subtracting background data from the raw data;
For the intermediate difference data output from the fixed pattern noise filter, a post-filter that reduces the random noise depending on the sensor position included in the sensor output data and outputs the difference data;
A touch determination unit that outputs touch determination data based on the difference data output from the subsequent filter;
A coordinate calculation unit that outputs the touched position coordinates based on the difference data output from the post-stage filter and the touch determination data output from the touch determination unit; A touch detection method applied to a touch detection device that outputs position coordinates,
In the fixed pattern noise filter,
An update step of updating the latest background data stored in the frame memory in response to a first control command from the outside;
A background data extraction step of reading out the latest background data stored in the frame memory;
A variable weighted averaging step of changing a weight for averaging the latest background data and the raw data read from the frame memory in accordance with an external second control command and outputting the background data after weighted averaging When,
A subtracting step of outputting the intermediate difference data by subtracting the background data after the weighted averaging output in the variable weighted averaging step from the raw data;
Based on the touch determination data output from the touch determination unit, a first generation for generating the first control command for storing the weighted average background data in the frame memory as the latest background data Steps,
A second generation step of generating the second control command for obtaining background data after the weighted averaging with a desired weight based on the touch determination data output from the touch determination unit; .

Images