Electronic device, information processing method and program

Abstract

Provided is an electronic device comprising: an event-driven vision sensor including a sensor array composed of sensors that generate an event signal upon detecting a change in the intensity of incident light; an inertial measurement unit (IMU) that has a known positional relationship with the vision sensor and measures at least one physical amount; and an additional information generation unit that generates additional information related to the event signal or an output value from the IMU by associating flow information detected on the basis of the event signal with a flow pattern that has been associated with the at least one physical amount or an estimation value based on the at least one physical amount in advance.

Description

[Technical Field]

[0001] The present invention relates to an electronic device, an information processing method, and a program. [Background technology]

[0002] Event-driven vision sensors are known in which pixels that detect changes in the intensity of incident light generate signals asynchronously. Event-driven vision sensors are advantageous in that they can operate at high speeds and with low power consumption compared to frame-type vision sensors, specifically CCD or CMOS image sensors, which scan all pixels at predetermined intervals. Technologies related to such event-driven vision sensors are described in, for example, Patent Document 1 and Patent Document 2. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Special Publication No. 2014-535098 [Patent Document 2] Japanese Patent Application Publication No. 2018-85725 Summary of the Invention [Problem to be solved by the invention]

[0004] However, although the advantages of event-driven vision sensors as described above are known, it cannot be said that sufficient peripheral technologies have yet been proposed that take into account characteristics that differ from those of conventional vision sensors, such as frame-type vision sensors.

[0005] Therefore, an object of the present invention is to provide an electronic device, an information processing method, and a program that enable useful information to be generated by utilizing an event-driven vision sensor and an inertial measurement unit (IMU). [Means for solving the problem]

[0006] According to one aspect of the present invention, there is provided an electronic device comprising: an event-driven vision sensor including a sensor array composed of sensors that generate an event signal when a change in the intensity of incident light is detected; an inertial measurement unit (IMU) whose positional relationship with the vision sensor is known and that measures at least one physical quantity; and an additional information generation unit that generates additional information related to the event signal or the output value of the IMU by correlating flow information detected based on the event signal with a flow pattern that is previously associated with the at least one physical quantity or an estimated value based on the at least one physical quantity.

[0007] According to another aspect of the present invention, there is provided an information processing method including a step of generating additional information related to the event signal or the output value of an inertial measurement unit (IMU) by correlating flow information detected based on an event signal received from an event-driven vision sensor including a sensor array configured with sensors that generate an event signal when a change in the intensity of incident light is detected with a flow pattern previously associated with at least one physical quantity measured by the IMU, whose positional relationship with the vision sensor is known, or an estimated value based on the at least one physical quantity.

[0008] According to yet another aspect of the present invention, there is provided a program for causing a computer to perform a function of generating additional information related to the event signal or the output value of an inertial measurement unit (IMU) by correlating flow information detected based on an event signal received from an event-driven vision sensor including a sensor array composed of sensors that generate an event signal when a change in the intensity of incident light is detected with a flow pattern previously associated with at least one physical quantity measured by the IMU, whose positional relationship with the vision sensor is known, or an estimated value based on the at least one physical quantity. [Brief explanation of the drawings]

[0009] [Figure 1]1 is a block diagram showing a schematic configuration of an electronic device according to a first embodiment of the present invention. [Figure 2] 2 is a block diagram showing the functional configuration of a processing circuit in the electronic device shown in FIG. 1. FIG. [Figure 3] FIG. 3 is a diagram illustrating an example of association between flow information and flow patterns in the first exemplary embodiment of the present invention. [Figure 4] FIG. 10 is a diagram illustrating an example of association between flow information and flow patterns in the second exemplary embodiment of the present invention. [Figure 5] FIG. 10 is a diagram illustrating an example of association between flow information and flow patterns in the second exemplary embodiment of the present invention. [Figure 6] FIG. 10 is a block diagram showing the functional configuration of a processing circuit in an electronic device according to a third embodiment of the present invention. [Figure 7] FIG. 10 is a diagram illustrating an example of a method for generating a flow pattern in the third embodiment of the present invention. [Figure 8] FIG. 11 is a diagram illustrating an example of association between flow information and flow patterns in the third exemplary embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, several embodiments of the present invention will be described in detail with reference to the accompanying drawings. In this specification and drawings, components having substantially the same functional configurations are designated by the same reference numerals, and redundant description will be omitted.

[0011] (First embodiment) 1 is a block diagram showing a schematic configuration of an electronic device according to a first embodiment of the present invention. As shown in FIG. 1, the electronic device 10 includes a vision sensor 100, a control unit 200, and an IMU 300.

[0012] The event-driven vision sensor 100 includes a sensor array 110 composed of sensors 110A, 110B, ... corresponding to pixels of an image, and a processing circuit 120 connected to the sensor array 110. The sensors 110A, 110B, ... include light-receiving elements and generate an event signal when they detect a change in the intensity of incident light, more specifically, a change in luminance. The event signal is output from the processing circuit 120 as information indicating, for example, a timestamp, sensor identification information (e.g., pixel position), and the polarity of the luminance change (increase or decrease). When an object moves within the field of view of the sensor array 110, the intensity of light reflected or scattered by the object changes. Therefore, the movement of the object can be detected in chronological order based on the event signals generated by the sensors 110A, 110B, ... corresponding to the edges of the object.

[0013] The control unit 200 includes a communication interface 210, a processing circuit 220, and a memory 230. The communication interface 210 receives event signals transmitted from the processing circuit 120 of the vision sensor 100 and output values ​​such as acceleration and angular velocity transmitted from the IMU 300, and outputs them to the processing circuit 220. The processing circuit 220 operates, for example, according to a program stored in the memory 230, and processes the received event signals and output values. For example, the processing circuit 220 generates images in time series that map positions where luminance changes occur based on the event signals, and temporarily or permanently stores the images in the memory 230 or transmits them to another device via the communication interface 210. Furthermore, for example, the processing circuit 220 estimates the attitude angle and displacement of the electronic device 10 by integrating the acceleration and angular velocity obtained as output values ​​from the IMU 300.

[0014] The IMU 300 (inertial measurement unit) measures at least one physical quantity. More specifically, for example, the IMU 300 includes a gyro sensor and an acceleration sensor and detects the angular velocity and acceleration occurring in the electronic device 10. The angular velocity and acceleration are examples of physical quantities. Here, the positional relationship between the IMU 300 and the vision sensor 100 is known. Specifically, for example, the attitude angle and displacement estimated based on the output value of the IMU 300 can be considered to be the same as the attitude angle and displacement of the vision sensor 100, or can be converted to the attitude angle and displacement of the vision sensor 100 by geometric calculation. When a displacement or rotation occurs in the electronic device 10, the IMU 300 detects the angular velocity and acceleration, and the positional relationship between the sensors 110A, 110B, etc. and the object changes, causing the sensors 110A, 110B, etc. corresponding to the object's edges to generate event signals. In other words, when a displacement or rotation occurs in the electronic device 10, the vision sensor 100 also generates an event signal reflecting the displacement or rotation.

[0015] Fig. 2 is a block diagram showing the functional configuration of a processing circuit in the electronic device shown in Fig. 1. In the example shown in Fig. 2, processing circuit 220 of control unit 200 includes an event signal analysis unit 221, an output value analysis unit 222, a correction information generation unit 223, and a correction processing unit 224 as functions implemented by operating in accordance with a program stored in memory 230, for example. The function of each unit will be described below.

[0016] The event signal analysis unit 221 analyzes the event signal received from the vision sensor 100 via the communication interface 210. Specifically, for example, the event signal analysis unit 221 detects flow information indicating the movement of a subject by mapping the time-series brightness changes indicated by the event signal. Here, the flow information is, for example, a vector representation of the movement of a subject in each region obtained by dividing a pixel region according to a predetermined rule. The shape and size of the divided region in which the flow information is generated are not particularly limited, but the following description may exemplify flow information generated in a grid-shaped divided region including a predetermined number of pixels in two directions, x and y.

[0017] The output value analysis unit 222 analyzes output values ​​such as acceleration and angular velocity received from the IMU 300 via the communication interface 210. Specifically, for example, the output value analysis unit 222 calculates an estimated value of the displacement of the electronic device 10 by integrating the acceleration obtained as an output value of the IMU 300. The output value analysis unit 222 also calculates an estimated value of the attitude angle of the electronic device 10 by integrating the angular velocity obtained as an output value of the IMU 300. The calculated estimated values ​​231 of the displacement and attitude angle of the electronic device 10 are stored in the memory 230. The output values ​​232 of the acceleration and angular velocity received from the IMU 300 may also be stored in the memory 230.

[0018] The correction information generator 223 generates correction information for the acceleration and angular velocity of the output value 232 of the IMU 300 acquired in synchronization with the flow information by associating the flow information detected by the event signal analyzer 221 with a pre-prepared flow pattern 233. The flow pattern 233 is stored in, for example, the memory 230. In the following description, the output value 232 acquired in synchronization with the flow information refers to, for example, the output value 232 having a timestamp that at least partially corresponds to the event signal that is the basis of the flow information, and does not necessarily mean that the time when the flow information was detected and the time of the output value 232 completely match. In this embodiment, the correction information generator 223 extracts a flow pattern that matches the flow information, as in the example described below, and generates correction information based on physical quantities, specifically, acceleration and angular velocity, associated with the extracted flow pattern.

[0019] 3, the flow pattern 233 associates acceleration (Accel) and angular velocity (Gyro) with the flow pattern based on the results of a previously performed measurement. More specifically, the flow pattern 233 is generated by, for example, inducing displacement and rotation at known acceleration and angular velocity in the electronic device 10 and detecting flow information from the event signal generated by the vision sensor 100 at that time. If the flow information detected by the event signal analysis unit 221 matches any of the flow patterns 233, the correction information generation unit 223 generates correction information for the acceleration and angular velocity based on the difference between the acceleration and angular velocity associated with the flow pattern 233 and the acceleration and angular velocity in the output value 232. The correction processing unit 224 corrects the acceleration and angular velocity in the output value 232 according to the generated correction information.

[0020] 3, whether flow information F matches flow pattern 233 may be determined using a cost function H(F). In this example, cost function H(F) indicates the magnitude of the difference between flow information F and each flow pattern, and a flow pattern with a smaller cost function H(F) is selected. The correction amounts of acceleration and angular velocity in the correction information may be set, for example, at a rate according to the value of cost function H(F) for the selected flow pattern, so that the acceleration and angular velocity in output value 232 approach the acceleration and angular velocity associated with flow pattern 233. In this case, the smaller the value of cost function H(F), the closer the acceleration and angular velocity in output value 232 approach the acceleration and angular velocity associated with flow pattern 233.

[0021] According to the configuration of the first embodiment of the present invention as described above, by previously associating the acceleration and angular velocity of the electronic device 10 with flow information detected from the event signal of the vision sensor 100, it is possible to correct the acceleration and angular velocity in the output value of the IMU 300 based on the flow information, thereby improving accuracy. The correction information for the acceleration and angular velocity is an example of additional information related to the output value of the IMU 300, and the correction information generator 223 is an example of an additional information generator that generates additional information by associating the flow information detected based on the event signal with a flow pattern.

[0022] (Second embodiment) Next, a second embodiment of the present invention will be described. In this embodiment, the functional configuration of the electronic device and processing circuit is the same as in the first embodiment. The difference from the first embodiment is that in this embodiment, as in the examples shown in FIGS. 4 and 5, the amount of change in velocity, acceleration, or attitude angle is associated with a pattern of flow information in flow pattern 233, and correction information generation unit 223 associates flow information with flow pattern 233 in a manner different from that in the first embodiment. In other respects, the configuration of this embodiment is the same as in the first embodiment, and therefore a detailed description thereof will be omitted.

[0023] FIG. 4 shows an example of a flow pattern associated with velocity and acceleration, and FIG. 5 shows examples of flow patterns (Pattern 1 to Pattern 3) associated with changes in attitude angle (Pan, Tilt, Roll). In this embodiment, the flow pattern 233 is generated by, for example, inducing a displacement at a known velocity and acceleration, or a rotation at a known angular velocity, in the electronic device 10, and detecting flow information from an event signal generated by the vision sensor 100 at that time. The velocity and acceleration correction values ​​can be used as correction values ​​when the output value analysis unit 222 calculates an estimated value of the displacement of the electronic device 10. The attitude angle correction value can be used as a correction value when the output value analysis unit 222 calculates an estimated value of the attitude angle of the electronic device 10.

[0024] In this embodiment, the correction information generating unit 223 generates correction information for the velocity, acceleration, or attitude angle based on the contribution rate of each flow pattern 233 when the flow information F detected by the event signal analyzing unit 221 is decomposed into components of a plurality of flow patterns 233 as in the above example. Specifically, the correction information generating unit 223 generates correction information by adding up the velocity, acceleration, or attitude angle associated with each of the plurality of flow patterns 233 according to the contribution rate of each flow pattern. For example, in the example of FIG. 4, the correction value C for the velocity v and acceleration correction value C a is calculated as the sum of the speeds v1, v2, and v3 and accelerations a1, a2, and a3 associated with each flow pattern multiplied by a coefficient corresponding to the contribution rate. Similarly, in the example of Figure 5, the correction value C p ,C t ,C ris calculated as the sum of the attitude angles associated with the respective flow patterns multiplied by a coefficient corresponding to the contribution rate. Strictly speaking, it is the amount of change in the attitude angle that is detected from the flow information F, but for example, by setting the time point at which it is determined that the electronic device 10 has come to a standstill based on the flow signal F as a reference point, the amount of change in the attitude angle detected from the flow information F between the reference points can be compared with the amount of change in the attitude angle measured by the IMU 300 to calculate a correction value for the attitude angle.

[0025] According to the configuration of the second embodiment of the present invention as described above, by previously associating the velocity, acceleration, or attitude angle of the electronic device 10 with flow information detected from the event signal of the vision sensor 100, it is possible to correct the attitude angle and displacement in the estimated values ​​based on the output value of the IMU 300 based on the flow information, thereby improving accuracy. The correction information for the estimated values ​​of the attitude angle and displacement is an example of additional information related to the output value of the IMU 300, and the correction information generator 223 is an example of an additional information generator that generates additional information by associating the flow information detected based on the event signal with a flow pattern.

[0026] In the second embodiment described above, correction information is generated by adding up values ​​associated with multiple flow patterns according to their contribution rates, so that valid correction information can be generated even when, for example, there is no flow pattern with a sufficiently small value of the cost function (H) in the first embodiment. In the case of the first embodiment, the more flow patterns there are to be compared, the higher the possibility of selecting a valid flow pattern. However, in cases where, for example, there are many variations in the movements and environmental conditions of the electronic device 10 and it is difficult to prepare flow patterns that cover all of them, the configuration of the second embodiment may be effective.

[0027] (Third embodiment) Fig. 6 is a block diagram showing the functional configuration of a processing circuit in an electronic device according to a third embodiment of the present invention. In the example shown in Fig. 6, the processing circuit 220 of the control unit 200 includes an event signal analysis unit 221, an output value analysis unit 222, and a depth information generation unit 225 as functions implemented by operating in accordance with a program stored in, for example, a memory 230. Note that the configuration of this embodiment is similar to that of the first embodiment except for the depth information generation unit 225, and therefore a detailed description thereof will be omitted.

[0028] The depth information generator 225 generates depth information of the subject of the vision sensor 100 by associating the flow information detected by the event signal analyzer 221 and the amount of change in attitude angle of the electronic device 10 based on the output value of the IMU 300 acquired in synchronization with the flow information with a flow pattern 234. The flow pattern 234 associates the amount of change in attitude angle and the depth of the subject with the flow pattern based on the results of measurements performed in advance. More specifically, the flow pattern 234 is generated by causing the electronic device 10 to change the attitude angle when the depth d of the subject 501 is known, as in the example shown in FIG. 7, and detecting flow information from the event signal generated by the vision sensor 100 at that time. Flow information is detected while causing various amounts of change in attitude angle while maintaining the depth d, and then flow information is similarly detected by changing the depth d. This allows the flow pattern 234 to be obtained, from which the depth d can be identified using the amount of change in attitude angle and the pattern of flow information as keys.

[0029] In this embodiment, the depth information generation unit 225 extracts flow patterns 234 corresponding to changes in posture angle based on output values ​​of the IMU 300 acquired in synchronization with the flow information, and identifies a pattern among the extracted flow patterns 234 to which the flow information at least partially matches. In the example shown in FIG. 8 , flow patterns (Pattern 1 to Pattern 3) corresponding to changes in posture angle (ΔPan, ΔTilt, ΔRoll) estimated from the output values ​​of the IMU 300 are extracted, and a pattern matching each of the divided regions of the flow information F is identified. In the illustrated example, Pattern 1 (depth d1) is identified as a matching pattern in the upper left portion of the pixel region, Pattern 2 (depth d2) is identified as a matching pattern in the upper right portion of the pixel region, and Pattern 3 (depth d3) is identified as a matching pattern in the lower portion of the pixel region. The depth information generation unit 225 generates depth information that identifies the depth of the subject in each portion of the pixel region based on the depths d1 to d3 of the subject associated with these flow patterns, respectively.

[0030] According to the configuration of the third embodiment of the present invention as described above, the depth d of the subject and flow information are previously associated with each amount of change in the attitude angle of the electronic device 10, thereby making it possible to generate depth information of the subject for the event signal. Note that, since the flow information changes depending on the influence of both the depth d of the subject and the amount of change in the attitude angle of the electronic device 10, the depth of the subject can be more appropriately identified by using the amount of change in the attitude angle estimated from the output value of the IMU 300. The depth information that identifies the depth of the subject is an example of additional information related to the event signal, and the depth information generation unit 225 is an example of an additional information generation unit that generates additional information by associating flow information detected based on the event signal with a flow pattern.

[0031] Although several embodiments of the present invention have been described in detail above with reference to the accompanying drawings, the present invention is not limited to these examples. It is clear that a person skilled in the art to which the present invention pertains can conceive of various modifications and alterations within the scope of the technical idea described in the claims, and it is understood that these also naturally fall within the technical scope of the present invention. [Explanation of symbols]

[0032] 10...electronic device, 100...vision sensor, 110...sensor array, 110A...sensor, 110B...sensor, 120...processing circuit, 200...control unit, 210...communication interface, 220...processing circuit, 221...event signal analysis unit, 222...output value analysis unit, 223...correction information generation unit, 224...correction processing unit, 225...depth information generation unit, 230...memory, 231...estimated value, 232...output value, 233...flow pattern, 234...flow pattern, 501...subject

Claims

1. an event-driven vision sensor including a sensor array configured with sensors that generate an event signal when detecting a change in the intensity of incident light; an inertial measurement unit (IMU) whose positional relationship with the vision sensor is known and which measures at least one physical quantity; an additional information generating unit that generates additional information related to the event signal or the output value of the IMU by associating flow information detected based on the event signal with a flow pattern that is previously associated with the at least one physical quantity or an estimated value based on the at least one physical quantity; An electronic device comprising:

2. The electronic device according to claim 1 , wherein the additional information includes correction information for the output value acquired in synchronization with the flow information or correction information for an estimated value based on the output value.

3. 3. The electronic device according to claim 2, wherein the additional information generation unit extracts the flow pattern that matches the flow information, and generates the correction information based on the at least one physical quantity associated with the extracted flow pattern or an estimated value based on the at least one physical quantity.

4. the flow patterns include first and second flow patterns; 3. The electronic device according to claim 2, wherein the additional information generation unit generates the correction information by adding together the at least one physical quantity associated with each of the first and second flow patterns or an estimated value based on the at least one physical quantity according to contribution rates of the first and second flow patterns when the flow information is decomposed into components of the first and second flow patterns.

5. the additional information includes depth information of the subject of the vision sensor; the flow pattern is associated with the depth of the object in addition to the at least one physical quantity or an estimate based on the at least one physical quantity; 5. The electronic device according to claim 1, wherein the additional information generation unit extracts the flow pattern corresponding to the output value acquired in synchronization with the flow information or an estimated value based on the output value, and generates the depth information based on a depth of the subject associated with a flow pattern to which the flow information at least partially matches among the extracted flow patterns.

6. An information processing method comprising: a step of generating additional information related to an output value of an inertial measurement unit (IMU) by correlating flow information detected based on an event signal received from an event-driven vision sensor including a sensor array composed of sensors that generate an event signal when a change in the intensity of incident light is detected with a flow pattern previously associated with at least one physical quantity measured by the IMU, the IMU having a known positional relationship with the vision sensor, or an estimated value based on the at least one physical quantity.

7. A program for enabling a computer to perform the function of generating additional information related to the event signal or the output value of an inertial measurement unit (IMU) by correlating flow information detected based on an event signal received from an event-driven vision sensor including a sensor array composed of sensors that generate an event signal when a change in the intensity of incident light is detected with a flow pattern that is pre-associated with at least one physical quantity measured by the IMU, whose positional relationship with the vision sensor is known, or an estimated value based on the at least one physical quantity.

Images