Dynamic vision sensor
By introducing bias change circuits into the dynamic vision sensor to generate event signals, the problem that traditional dynamic vision sensors cannot output signals when the light intensity remains unchanged or changes are small, ensuring the vehicle detection capability under different lighting conditions.
Patent Information
- Application Number
- CN202410847826.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2023-07-13
- Filing Date
- 2024-06-27
- Publication Date
- 2025-01-14
AI Technical Summary
Traditional dynamic vision sensors only generate event signals when the incoming light intensity changes, and cannot output valid signals when the light intensity remains unchanged or changes, resulting in the incoming vehicle being unable to detect in some scenarios (such as when the rear vehicle is at the same speed as the front vehicle).
A dynamic vision sensor is designed to generate event signals artificially through a bias change circuit when the incoming light intensity remains unchanged or changes. It uses variable bias to output event signals to ensure that grayscale levels can be generated under different lighting conditions. image.
It realizes that event signals can be output even when the light intensity remains unchanged or changes are small, ensuring that the vehicle detection system operates normally under different lighting conditions, and enhancing the detection ability of the vehicles ahead.
Smart Images

Figure CN119316735A_ABST
Abstract
Description
[0001] This application claims the priority of Korean Patent Application No. 10-2023-0091355 filed on July 13, 2023 in the Korean Intellectual Property Office, the disclosure of which is incorporated herein by reference in its entirety. Technical Field
[0002] Various example embodiments described herein relate to a dynamic vision sensor (DVS), an image signal processor that operates based on an output of the dynamic vision sensor, and / or an application processor (eg, an application processor that includes the DVS). Background Art
[0003] An image sensor is a device that converts an optical signal into an electrical signal, or includes a device that converts an optical signal into an electrical signal, or is included in a device that converts an optical signal into an electrical signal, and may include and / or be implemented as one or more of a charge coupled device (CCD) image sensor, a complementary metal oxide semiconductor (CMOS) image sensor, and a DVS sensor.
[0004] The output value of a pixel of a CMOS image sensor may change according to the intensity of incident light. The CMOS image sensor generates an image signal based on the intensity of incident light. On the other hand, a DVS sensor detects an event in which the intensity of incident light changes and generates an event signal. Summary of the invention
[0005] Various example embodiments may provide a dynamic vision sensor capable of generating an event signal even when the intensity of incident light does not change.
[0006] Alternatively or additionally, various example embodiments may provide a processor capable of generating a grayscale image based on the time signal.
[0007] According to some example embodiments, a dynamic vision sensor (DVS) includes: a photodiode configured to generate a first photodiode current in response to a change in light due to movement of an object; a first transistor; a first node connected to the photodiode and the first transistor; a second node connected to the first transistor and configured to receive a first voltage and to receive a second voltage; a ground; a second transistor connected to the first node and the ground; a third node connected to the first transistor and the second transistor; and a current source connected to the third node, wherein the first voltage has a first level and the second voltage has a second level different from the first level.
[0008] According to some example embodiments, a dynamic vision sensor (DVS) includes: a photodiode configured to generate a first photodiode current in response to a change in light due to movement of an object; a first transistor; a first node connected to the photodiode and the first transistor; a second node connected to the first transistor; a ground; a second transistor connected to the first node and the ground; a third node connected to the first transistor and the second transistor; and a current source connected to the third node and configured to receive a first voltage and a second voltage, wherein the first voltage has a first level and the second voltage has a second level different from the first level. BRIEF DESCRIPTION OF THE DRAWINGS
[0009] The above and other objects and features of various exemplary embodiments will become apparent by describing in detail embodiments of various exemplary embodiments with reference to the attached drawings.
[0010] Figure 1 is a block diagram illustrating a dynamic vision sensor according to some example embodiments.
[0011] Figure 2 is a diagram showing gain characteristics of a dynamic vision sensor according to illumination.
[0012] Figure 3 yes Figure 1 A block diagram of an example of a dynamic vision sensor.
[0013] Figure 4 yes Figure 1 A block diagram of an example of a dynamic vision sensor.
[0014] Figure 5A is based on Figure 1 A circuit diagram of an embodiment of a dynamic vision sensor.
[0015] Figure 5B , Figure 5C , Figure 5D and Figure 5E is based on Figure 5A A circuit diagram of a photoreceptor of an embodiment.
[0016] Figure 6 is a diagram illustrating an example of a voltage buffer circuit of a dynamic vision sensor according to some example embodiments.
[0017] Figure 7 Show Figure 1 Schematic diagram of a method for dynamically varying the bias current of a vision sensor.
[0018] Figure 8 It is shown Figure 1 Schematic diagram of an example of a bias variation circuit for a dynamic vision sensor.
[0019] Fig. 9 It is shown Figure 1 A diagram of an example of a bias variation circuit for a dynamic vision sensor.
[0020] Fig.10 is a diagram illustrating an example of a photoreceptor of a dynamic vision sensor according to some example embodiments.
[0021] Fig.11 is a diagram illustrating an example of a photoreceptor of a dynamic vision sensor according to some example embodiments.
[0022] Fig.12 is a diagram illustrating an example of a photoreceptor of a dynamic vision sensor according to some example embodiments.
[0023] Fig.13 is a diagram illustrating a configuration of a plurality of substrates of a dynamic vision sensor according to some example embodiments.
[0024] FIG. 14A to FIG. 14C is a diagram illustrating a configuration of a plurality of substrates of a dynamic vision sensor according to some example embodiments.
[0025] Fig.15 is a block diagram illustrating a dynamic vision sensor and a processor according to some example embodiments.
[0026] Fig.16 are diagrams illustrating various examples of electronic devices to which a dynamic vision sensor and a processor are applied according to some example embodiments.
[0027] Fig.17 is a flowchart for describing a method in which a processor generates a grayscale image according to some example embodiments.
[0028] Fig.18 is a diagram for describing a method in which a processor generates a grayscale image according to some example embodiments.
[0029] Fig.19 is a diagram illustrating results of generating a grayscale image by varying different biases of a dynamic vision sensor according to some example embodiments. DETAILED DESCRIPTION
[0030] Hereinafter, various exemplary embodiments may be described in detail and clearly to such an extent that a person having ordinary skill in the art can easily implement the present disclosure.
[0031] Figure 1 is a block diagram illustrating a dynamic vision sensor according to some example embodiments. Figure 1 , the dynamic vision sensor 10 includes a pixel 100 and a bias variation circuit 200 .
[0032] The dynamic vision sensor 10 outputs an event signal ES generated based on a change in incident light (eg, a change in the intensity of the incident light).
[0033] The photoreceptor 110 of the pixel 100 includes a photoelectric conversion element and a converter circuit. FIG. 5A to FIG. 5E , the photoelectric conversion element may be or may include a photodiode PD; however, example embodiments are not limited thereto. Various example embodiments describe that the photoelectric conversion element is a photodiode PD. However, the photoelectric conversion element may alternatively or additionally be or include a photocapacitor, a photogate, a pinned photodiode PPD, and a partially pinned photodiode, or a combination thereof.
[0034] The photodiode outputs a photodiode current IPD based on incident light. The converter circuit converts the photodiode current IPD flowing through the photoelectric conversion element into a logarithmic scale voltage VLOG and outputs the logarithmic scale voltage VLOG. The converter circuit may be a logarithmic amplifier, or may include a logarithmic amplifier or be included in a logarithmic amplifier. As described herein, the output of the current does not indicate the direction of the current, and in some example embodiments may indicate or correspond to the magnitude of the current.
[0035] The logarithmic scale voltage VLOG of the photoreceptor 110 is provided to the pixel backend circuit. The pixel backend circuit includes a differential circuit 120 and a comparator 130. The pixel backend circuit may output an event signal ES "corresponding to or indicating a change in the intensity of the incident light" based on the logarithmic scale voltage VLOG. For example, when an event in which the intensity of the incident light increases occurs, the pixel backend circuit may output an event signal ES indicating an ON event. When an event in which the intensity of the incident light decreases occurs, the pixel backend circuit may output an event signal ES indicating an OFF event. The event signal ES may include a plurality of bits corresponding to polarity information indicating whether there is an ON event and an OFF event.
[0036] The dynamic vision sensor 10 may be configured to process events for each pixel or to process events in columns. For example, when processing events for each pixel, each time the corresponding pixel 100 outputs an event signal ES, the dynamic vision sensor 10 may notify the address event representation (AER) that the event occurs in the corresponding pixel 100. The column AER and row AER of the dynamic vision sensor 10 receive event signals including polarity information from the pixel 100 where the event occurs. When processing events in columns, the dynamic vision sensor 10 sequentially scans each column of the pixel array based on the clock timing. The corresponding pixel 100 of the selected column sends the event signal ES including polarity information to the row sampling circuit.
[0037] Even when the intensity of incident light does not change or changes slightly, the dynamic vision sensor 10 according to various exemplary embodiments may still output the event signal ES. The dynamic vision sensor 10 may generate the event signal ES by changing at least one of the biases to be supplied to the pixel. At least one of the biases to be supplied to the pixel may be changed by the bias changing circuit 200.
[0038] The bias change circuit 200 receives a bias change enable signal or a bias change disable signal from an external processor located outside the dynamic vision sensor 10. The bias change circuit 200 changes at least one of the biases to be supplied to the pixel in response to the bias change enable signal. For example, the bias change circuit 200 may change at least one of the multiple biases to be supplied to the photoreceptor 110.
[0039] The bias changing circuit 200 may change a selected bias among at least two bias levels at a specific period (such as a dynamically determined period, or alternatively, a predetermined period). For example, the bias changing circuit 200 may periodically change a power supply voltage to be supplied to the photoreceptor 110 among a first power supply voltage level and a second power supply voltage level.
[0040] Example embodiments have experimentally identified that the conversion gain of the converter circuit varies according to the magnitude of the photodiode current IPD of the dynamic vision sensor. Figure 2 (where the horizontal axis represents frequency (Hz) and the vertical axis represents the AC gain of the logarithmic scale voltage VLOG of 1 lux), it has been identified that the gain GA2 from the merged four pixels is greater than the gain GA1 from the non-merged pixels of the same size. In addition, it has been identified that the gain GA3 from the merged 16 pixels is greater than the gain GA2 from the merged 4 pixels.
[0041] Therefore, the conventional dynamic vision sensor generates an event signal only when the intensity of the incident light changes. Meanwhile, the dynamic vision sensor 10 according to various exemplary embodiments may output the event signal ES even when the intensity of the incident light does not change or changes very little. For example, the dynamic vision sensor 10 may periodically change the bias to be supplied to the photoreceptor 110 even when the intensity of the incident light does not change or changes very little. As a result, the photodiode current IPD changes. When the photodiode current IPD changes, the dynamic vision sensor 10 according to various exemplary embodiments may output the event signal ES even when the intensity of the incident light does not change or changes very little.
[0042] Even when the incident light intensity does not change or changes very little, the dynamic vision sensor 10 can output the event signal ES based on the changed bias rather than based on the change in the intensity of the incident light. The dynamic vision sensor 10 and / or the processor can generate a grayscale image by using the event signal ES based on the variable bias. Therefore, even when the intensity of the incident light does not change or changes very little, the normal operation of the application can be guaranteed or more likely to be guaranteed.
[0043] For example, when the rear vehicle is driving at the same speed as the front vehicle, an advanced driver assistance system (ADAS) in a vehicle based on a conventional dynamic vision sensor and / or an object detection system in an autonomous vehicle may not be able to detect the front vehicle. For example, when the rear vehicle is driving at the same speed as the front vehicle, the pixels of the conventional dynamic vision sensor may not be able to generate an event signal. Therefore, the object detection system of the rear vehicle may be temporarily and / or unable to recognize the front vehicle during a specific time. On the other hand, even when the rear vehicle is driving at the same speed as the front vehicle, the dynamic vision sensor 10 according to various example embodiments may artificially generate an event signal ES by changing the bias, and then a grayscale image may be generated. Therefore, the object detection system of the rear vehicle can generally recognize the front vehicle normally based on the grayscale image.
[0044] Figure 3 is a block diagram illustrating a dynamic vision sensor 10 according to some example embodiments. Figure 3 The dynamic vision sensor 10 includes a pixel array PXA, a bias variation circuit 200 , a column driver 300 , a row sampling circuit 400 , and a packetizer or packetizer circuit 500 .
[0045] The pixel array PXA includes a plurality of pixels 100. For example, the plurality of pixels 100 may be arranged in a matrix form, such as a rectangular matrix (eg, a square matrix form). There may be dummy pixels (not shown) and / or redundant pixels (not shown); example embodiments are not limited thereto.
[0046] The column driver 300 selects one of the multiple columns of the pixel array PXA. The column driver 300 can select a column by sending selection signals CLn to CL0. The column driver 300 can sequentially select multiple columns with a specific dynamically determined cycle or a predetermined cycle. For example, the column driver 300 can select the first column of pixels by setting the first column selection signal CLn to a high level based on the synchronization signal. After the first column of the pixel sends the event signal ESm to ES0 to the row sampling circuit 400, the column driver 300 can set the second column selection signal CLn-1 to a high level based on the synchronization signal. In other words, the dynamic visual sensor 10 can process the event signal of each pixel in a column unit based on the synchronization signal. Here, m and n are integers greater than or equal to 0, and n can be greater than, less than or equal to m.
[0047] The pixels of the column selected by the column driver 300 transmit the event signals ESm to ES0 to the row sampling circuit 400. The event signal may include a plurality of bits corresponding to polarity information indicating whether there is an on event and an off event. The row sampling circuit 400 may transmit a reset signal RS (e.g., RSm to RS0) to the pixel transmitting the event signal (on event signal) including the on event. The pixel 100 receiving the reset signal RS may initialize the differential circuit 120.
[0048] The row sampling circuit 400 may generate an address of the pixel 100 that “generates an on event signal”, and then transmit the address and the event signals ESm to ES0 to the packetizer 500. Alternatively, the row sampling circuit 400 may transmit the addresses and the event signals ESm to ES0 of all pixels 100 of the selected column to the packetizer 500. The address may include at least one of the column address and the row address of the pixel 100. The row sampling circuit 400 may transmit the time when the on event occurs to the packetizer 500. When only the address and the event signals ESm to ES0 of the pixel 100 that generates the on event signal are transmitted to the packetizer 500, the row sampling circuit 400 may transmit the time when the on event occurs to the packetizer 500 together.
[0049] The packetizer 500 generates and outputs event data EDT obtained by packetizing event signals ESm to ES0, addresses, and the time when the on event occurs, for the pixel 100 generating the on event signal or all pixels 100 in a specific column. The packet may be generated in units of pixels, columns, or frames including all columns. The packetizer 500 sends the generated packet to the processor.
[0050] The bias change circuit 200 of the dynamic visual sensor 10 according to some example embodiments may receive a bias change enable signal. The bias change controller 210 may periodically change the bias data and send the bias data to the bias generator 220. The bias generator 220 may provide the changed bias to the corresponding pixel 100. The bias change controller 210 may periodically change the bias among a plurality of bias levels. At least one bias included in the photoreceptor 110 of the pixel 100 may be changed based on the changed bias. The pixel 100 may generate event signals ESm to ES0 based on the variable bias, and the event signals ESm to ES0 may be sent to the row sampling circuit 400.
[0051] The bias change circuit 200 may transmit a bias change ON signal corresponding to the bias change enable signal or a bias change OFF signal corresponding to the bias change disable signal to the packetizer 500. The packetizer 500 may generate a packet (referred to as "event data") to which "information bits corresponding to the bias change ON signal or the bias change OFF signal" are added, and may transmit the packet to the processor. Therefore, the processor may determine whether the event signal included in the received packet is an event signal generated by a change in the intensity of incident light or an event signal generated based on a changed bias.
[0052] Figure 4 is a block diagram showing a dynamic vision sensor 10 according to some example embodiments. Figures 1 to 3 The same or similar components as those described above will be marked by using the same or similar reference numerals, and redundant descriptions will be omitted.
[0053] Reference Figure 4 The dynamic vision sensor 10 includes a pixel array PXA, a bias variation circuit 200, a column AER 300_1, a row AER 400_1 and a packetizer 500_1.
[0054] The pixel array PXA includes a plurality of pixels 100. For example, the plurality of pixels 100 may be arranged in a matrix form.
[0055] The pixel 100 may transmit an occurrence notification CR of an on event to the column AER 300_1. The column AER 300_1 may generate a column address of a pixel in which an on event has occurred. In response to the occurrence notification CR of the on event, the column AER 300_1 may transmit an acknowledgement signal Ack to the pixel in which the on event has occurred. The pixel 100 receiving the acknowledgement signal Ack may transmit an event signal ES to the row AER 400_1. The row AER 400_1 may transmit a reset signal RS to the pixel 100 transmitting the event signal ES in response to the event signal ES. The pixel 100 receiving the reset signal RS may initialize the differential circuit 120.
[0056] The row AER 400_1 may generate a row address of the pixel 100 that "generates an on event signal". The time stamper 400_2 may generate a time stamp TS indicating a time point at which an on event occurs. The row AER 400_1 may send the row address, column address, time stamp TS and event signal ES to the packetizer 500_1.
[0057] The packetizer 500_1 may generate and output event data EDT obtained by packetizing the row address, column address, time stamp TS, and event signal ES of the pixel 100 that “generates the on event signal.” The packetizer 500_1 may transmit the generated packet to the processor.
[0058] In the dynamic visual sensor 10 according to some example embodiments, the bias change circuit 200 may receive the bias variable enable signal and may change at least one bias included in the photoreceptor 110 of the pixel 100. The pixel 100 may notify the column AER 300_1 of an on event generated based on the changed bias. The pixel 100 may generate an event signal ES based on the changed bias and may transmit the event signal ES to the row AER 400_1.
[0059] The bias change circuit 200 may send a bias change on signal corresponding to the bias change enable signal or a bias change off signal corresponding to the bias change disable signal to the packetizer 500_1. The packetizer 500_1 may generate a packet by adding an information bit corresponding to the bias change on signal or the bias change off signal together with the row address, the column address, the time stamp TS, and the event signal ES. The packetizer 500_1 may send the packet to which the “information bit corresponding to the bias change on signal or the bias change off signal” is added to the processor. Therefore, the processor may determine whether the event signal included in the received packet is an event signal generated by a change in the intensity of the incident light or an event signal generated based on the changed bias.
[0060] Figure 5A It shows that according to Figure 3 A circuit diagram of a dynamic vision sensor 10 of an embodiment. Figure 5B , Figure 5C , Figure 5D and Figure 5E It shows that according to Figure 5A Components identical or similar to the above components will be marked by using the same or similar reference numerals, and redundant descriptions will be omitted. The dynamic vision sensor 10 in FIG. 5 is configured according to Figure 3 The embodiment of the dynamic vision sensor 10 is described under the premise of the pixel 100. However, according to Figure 4 The pixels 100 of the dynamic vision sensor 10 of the embodiment may also be implemented similarly.
[0061] Reference Figure 5A , a dynamic vision sensor 10 according to some example embodiments includes a photoreceptor 110 , a differentiation circuit 120 , a comparator 130 , an output logic 140 , and a bias variation circuit 200 .
[0062] The photoreceptor 110 includes a photodiode PD, a feedback transistor MFBN, and a converter circuit. The converter circuit may be a logarithmic amplifier circuit.
[0063] The logarithmic amplifier circuit may include a logarithmic amplifier transistor MAMP connected to the gate node of the feedback transistor MFBN and a constant current source 111. For example, the logarithmic amplifier transistor MAMP may be connected to the first node N1 and ground. The gate node of the logarithmic amplifier transistor MAMP may be connected to the photodiode PD.
[0064] The photodiode voltage VPD is applied to the node N1, and the photodiode current IPD flows through the photodiode PD. The photodiode current IPD may be different based on the level of the power supply voltage to the node N2. In this specification, the output of the current does not indicate the direction of the current. The logarithmic amplifier circuit converts the photodiode current IPD into a logarithmic scale voltage VLOG. The logarithmic scale voltage VLOG is provided to the differential circuit 120 of the pixel back-end circuit through the node N4.
[0065] The differential circuit 120 may receive the logarithmic scale voltage VLOG from the photoreceptor 110 and may generate an output voltage VDIFF obtained by amplifying the logarithmic scale voltage VLOG. In another embodiment, the differential circuit 120 may selectively generate an output voltage VDIFF obtained by amplifying the logarithmic scale voltage VLOG. Figure 6 The voltage buffer circuit 150 receives the voltage VSF and obtains the output voltage VDIFF. The voltage buffer circuit 150 (eg, the transistor SF of the voltage buffer circuit 150) may receive the logarithmic scale voltage VLOG from the photoreceptor 110 and output the voltage VSF (eg, through the node N5 connected to the node N4).
[0066] The differential circuit 120 includes a first capacitor C1, a second capacitor C2, an operational amplifier (-A) 121, and a reset switch RST_SW. When the differential circuit 120 generates an output voltage VDIFF obtained by amplifying the voltage VSF provided by the voltage buffer circuit 150, the capacitance ratio or capacitance value of the first capacitor C1 and the second capacitor C2 may be different from the capacitance ratio or capacitance value in the case of receiving the logarithmic scale voltage VLOG.
[0067] The differential circuit 120 is connected to the output node N3 of the logarithmic amplifier circuit of the photoreceptor 110 or the voltage buffer circuit 150 through the first capacitor C1. The second capacitor C2 is connected between the input and output of the operational amplifier 121. The reset switch RST_SW is connected in parallel with the second capacitor C2 between the input and output of the operational amplifier 121. The reset switch RST_SW is closed in response to the reset signal RS, and the differential circuit 120 is initialized. The reset signal RS may be received from the output logic 140.
[0068] Differentiating circuit 120 amplifies input voltage received from logarithmic amplifier circuit or voltage buffer circuit 150 according to a capacitance ratio between first capacitor C1 and second capacitor C2. The capacitance of each of first capacitor C1 and second capacitor C2 may be selected in consideration of the amplification ratio and a time interval of consecutive event occurrences.
[0069] The comparator 130 compares the level of the output voltage VDIFF of the differential circuit 120 with the level of the reference voltage VREF thus provided. Based on the comparison result, the comparator 130 outputs an event signal for determining whether the intensity of light incident on the photodiode PD increases or decreases. The comparator 130 may include a plurality of comparison circuits COMP1 to COMP2. When the intensity of the incident light increases, the first comparison circuit COMP1 in the comparator 130 may output a voltage of a high level, and the second comparison circuit COMP2 in the comparator 130 may output a voltage of a low level. When the intensity of the incident light decreases, the first comparison circuit COMP1 in the comparator 130 may output a voltage of a low level, and the second comparison circuit COMP2 in the comparator 130 may output a voltage of a high level. The plurality of comparison circuits COMP1 and COMP2 may have different reference voltages for determining an on event signal and an off event signal.
[0070] The output logic 140 may output the event signals ON_OUT and OFF_OUT. The output logic 140 may provide the event signals ON_OUT and OFF_OUT to the reference Figure 3 The row sampling circuit 400 described or referred to Figure 4 The output logic 140 may respond to the on-event signal ON_OUT from Figure 3 The row sampling circuit 400 shown in Figure 4 The row AER400_1 shown in FIG. 4 receives the reset signal RS.
[0071] Even when the intensity of incident light does not change, the dynamic vision sensor 10 according to some example embodiments may output event signals ON_OUT and OFF_OUT through the changed bias provided by the bias changing circuit 200 .
[0072] The bias varying circuit 200 may vary at least one bias among a plurality of biases provided to the photoreceptor 110. For example, the bias varying circuit 200 may vary at least one of a power supply voltage bias VB1 provided to the photoreceptor 110 and a bias current IBIAS of the logarithmic amplifier circuit.
[0073] In one embodiment, when a bias variation disable signal is received, the bias variation circuit 200 provides the power supply voltage VDD to the photoreceptor 110 through the node N2. When a bias variation enable signal is received, the bias variation circuit 200 may provide the first level of the power supply voltage bias VB1 and the second level of the power supply voltage bias VB1 to the photoreceptor 110 by periodically varying the power supply voltage bias VB1. The level of the power supply voltage VDD may be different from the level of the power supply voltage bias VB1. Figure 5B , when the power supply voltage VDD is supplied to the node N2 so that the photodiode voltage VPD1 is supplied to the node N1, the first photodiode current IPD1 flows through the photodiode PD in response to the change of light caused by the movement of the object. The logarithmic amplifier circuit can convert the first photodiode current IPD1 into a logarithmic scale voltage VLOG1. Figure 5C When the first level power supply voltage bias VB1 is provided to the node N2 so that the photodiode voltage VPD2 is provided to the node N1, the second photodiode current IPD2 flows through the photodiode PD. The logarithmic amplifier circuit can convert the second photodiode current IPD2 into a logarithmic scale voltage VLOG2. Figure 5D When the second level power supply voltage bias VB1 is provided to the node N2 so that the photodiode voltage VPD3 is provided to the node N1, the third photodiode current IPD3 flows through the photodiode PD. The logarithmic amplifier circuit can convert the third photodiode current IPD3 into a logarithmic scale voltage VLOG3.
[0074] In one embodiment, when the bias variation disable signal is received, the bias variation circuit 200 provides the first level of power supply voltage bias VB1 to the photoreceptor 110 through the node N2. The first level of power supply voltage bias VB1 may be the same as the power supply voltage VDD. When the bias variation enable signal is received, the bias variation circuit 200 may provide the photoreceptor 110 with the power supply voltage bias VB1 obtained by periodically varying the first level of power supply voltage and the second level of power supply voltage different from the first level.
[0075] In one embodiment, referring to Figure 5EThe switch SW may select the power supply voltage provided to the node N2 from the power supply voltage bias VB1 and the power supply voltage VDD based on the bias change enable signal and the bias change disable signal. The bias change circuit 200_1 may provide the power supply voltage bias VB1 based on the bias change enable signal.
[0076] Alternatively, in one embodiment, when a bias variation disable signal is received, the bias variation circuit 200 provides a voltage bias VB2 to the constant current source 111 of the photoreceptor 110, so that the photodiode PD generates a fourth photodiode current and the bias current of the third level flows to the node N3. When a bias variation enable signal is received, the bias variation circuit 200 may periodically vary the voltage bias VB2 provided to the constant current source 111, so that the photodiode PD generates a fifth photodiode current and a sixth photodiode current and the bias current of the third level and the bias current of the fourth level lower than the third level periodically change and flow. In another embodiment, when a bias variation enable signal is received, the bias variation circuit 200 may periodically vary the voltage bias VB2 provided to the constant current source 111, so that the bias current of the fourth level and the bias current of the fifth level periodically change and flow. The bias current of the fourth level and the bias current of the fifth level may be different from the bias current of the third level. For example, the photodiode PD may alternately receive the fourth level power supply voltage bias VB2 and the fifth level power supply voltage bias VB2.
[0077] For example, refer to Figure 7 , the constant current source 111 of the logarithmic amplifier circuit may be a constant current source circuit 111_1 in which a plurality of transistors MB2_1 and MB2_2 are connected in series to a power supply voltage VDD. When the bias variation circuit 200 receives a bias variation disable signal, the bias current of the third level flows to the node N3 by setting the voltage biases VB2_1 and VB2_2 of the gate nodes of the transistors MB2_1 and MB2_2 of the constant current source circuit 111_1. When the bias variation enable signal is received, the bias variation circuit 200 may periodically vary the voltage biases VB2_1 and VB2_2 of the gate nodes of the transistors MB2_1 and MB2_2. Therefore, the magnitude of the bias current IBIAS may periodically vary between the third level and a fourth level less than the third level.
[0078] Figure 8 2 is a block diagram showing a configuration of a bias varying circuit 200 according to some example embodiments. The bias varying circuit 200 includes a bias varying controller 210 and a bias generator 220.
[0079] Reference Figure 8, the bias generator 220 may include a plurality of bias circuits Bias 1 to Bias N capable of outputting a plurality of biases VOUT1 to VOUTN. The bias generator 220 may receive a bias address Bias Addr indicating a bias to be set and bias data Bias Data indicating a value of the bias to be set from the bias variation controller 210 .
[0080] The selector 221 of the bias generator 220 may latch the received bias address Bias Addr and the received bias data Bias Data. The decoder 222 may select the bias to be changed by decoding the received bias address Bias Addr. The selector 221 may set the value corresponding to the received bias data Bias Data as the setting bit of the circuit 223 of the selected bias. For example, the bias address Bias Addr may be decoded, and the gate node voltage of the first transistor included in the bias circuit Bias 1 corresponding to the bias address Bias Addr may be selected as the bias to be set. The bias voltage value based on the bias data Bias Data may be set as the voltage value of the gate node voltage of the first transistor. The gate node voltage of the first transistor may be output as the setting voltage value (VOUT1).
[0081] In response to receiving the bias change disable signal, the bias change controller 210 according to some example embodiments may send the first bias data Bias Data 1 to the bias generator 220 so as to output a bias having a first value. The bias change controller 210 may periodically and alternately send the first bias data BiasData 1 and the second bias data Bias Data 2 to the bias generator 220 in response to receiving the bias change enable signal. The first bias data Bias Data 1 and the second bias data Bias Data 2 may be sent alternately based on the clock signal Clock. The second bias data Bias Data 2 may be bias data that “sets the bias generator to output a bias having a value lower than the first value”.
[0082] The bias address Bias Addr may be pre-set for the bias change controller 210. The bias address Bias Addr corresponding to one or more biases may be pre-stored in a register. The bias data Bias Data, which is a value to be changed for each bias, may be pre-stored in a register. The bias change controller 210 may receive an information bit indicating a specific bias, and may select the bias address Bias Addr and the bias data BiasData of the bias corresponding to the information bit from the register. Optionally, the bias address Bias Addr and / or the bias data Bias Data may be received together with a bias change enable / disable signal.
[0083] Fig. 9 2 is a block diagram showing a configuration of a bias variation circuit 200 according to some example embodiments. The bias variation circuit 200 includes a bias variation controller 210_1 and a bias generator 220. Figure 8 Detailed descriptions of parts that are the same or similar to those described above will be omitted to avoid redundancy.
[0084] The bias variation controller 210_1 according to some example embodiments may pre-store bias addresses BiasAddr and bias data Bias Data of a plurality of biases in the register 210_11.
[0085] The bias variation controller 210_1 may receive a bias variation enable / disable signal. The bias variation controller 210_1 may receive an information bit for specifying a bias.
[0086] The bias change controller 210_1 may receive an information bit for specifying one of a plurality of bias change modes. For example, the bias change mode applied to a dynamic vision sensor used in a front camera of a vehicle may be different from that of a dynamic vision sensor used in an in-vehicle camera of a vehicle. For example, a dynamic vision sensor used in a front camera of a vehicle may change a power supply voltage of a photoreceptor in a bias change mode. On the other hand, a dynamic vision sensor used in an in-vehicle camera of a vehicle may change a bias current of a photoreceptor in a bias change mode.
[0087] The bias change selector 210_12 may send a bias address Bias Addr of a bias corresponding to the information bit to the bias change commander 210_13. The bias change commander 210_13 may select bias data Bias Data corresponding to the bias address Bias Addr and may provide the selected bias data Bias Data to the bias generator 220 together with the bias address Bias Addr.
[0088] Fig.10 is a circuit diagram of a photoreceptor 110_1 according to some example embodiments. Fig.10 In the photoreceptor 110_1 in FIG. 1 , the logarithmic amplifier circuit converts and outputs a bipolar junction transistor (BJT) current IBJT flowing into a BJT element BJTPD connected to a photodiode PD (rather than a photodiode current IPD flowing into the photodiode PD) into a logarithmic scale voltage VLOG. Detailed descriptions of the same or similar parts as those described above will be omitted to avoid redundancy.
[0089] Reference Fig.10, the photodiode PD is connected to the base of the BJT element BJTPD, and is connected to the power supply voltage VDD through a resistor 112 through which a current IR flows. The base current IB is amplified, and the collector current flows into the collector of the BJT element BJTPD. The current obtained by summing the collector current and the base current flows into the emitter of the BJT element BJTPD. The BJT current IBJT flows with a current greater than the photodiode current IPD (IBJT=(β+1)IPD; β=collector current / base current). Therefore, the drag phenomenon of the object caused by the reduction of the current flowing into the photodiode of the conventional dynamic vision sensor can be reduced in a low light environment. In addition, according to Fig.10 The dynamic vision sensor of embodiments may reduce the time required to generate events in low-light environments.
[0090] The pixel back-end circuit generates an event signal based on the logarithmic-scale voltage VLOG obtained by converting the BJT current IBJT.
[0091] The bias variation circuit 200 may vary at least one bias supplied to the photoreceptor 110_1. For example, the power supply voltage bias VB1 supplied to the photoreceptor 110_1 or the bias current IBIAS flowing through the constant current source 111 of the logarithmic amplifier circuit may vary. When the bias current IBIAS flowing into the constant current source 111 is varied, the bias variation circuit 200 may vary the voltage bias VB2 supplied to the constant current source 111. Therefore, even when the intensity of incident light does not change or changes very little, the bias variation circuit 200 may vary according to the power supply voltage bias VB1 supplied to the photoreceptor 110_1. Fig.10 The dynamic vision sensor of the embodiment may also output the event signal ES based on the changed bias, rather than outputting the event signal ES based on the change in the intensity of the incident light. Fig.10 The dynamic vision sensor 10 or processor of the embodiment can generate a grayscale image by using the event signal ES based on the variable bias. Therefore, even when the intensity of the incident light does not change or changes very little, the normal operation of the application can be guaranteed.
[0092] Fig.11 is a circuit diagram of a photoreceptor 110_2 according to some example embodiments. FIG. 5A to FIG. 5E The photoreceptor 110 described is different, Fig.11 The photoreceptor 110_2 includes a feedback gain control transistor MFBP connected to the feedback transistor MFBN and the photodiode PD. Detailed description of parts identical or similar to the above parts will be omitted to avoid redundancy.
[0093] The drain node of the feedback gain control transistor MFBP according to some example embodiments is connected to the source node of the feedback transistor MFBN. The source node of the feedback gain control transistor MFBP is connected to the gate node of the photodiode PD and the logarithmic amplifier transistor MAMP through the node N1. The capacitance of the capacitor included in the differential circuit 120 can be reduced by adjusting the gain by the feedback gain control transistor MFBP. Therefore, the size of the differential circuit 120 can be reduced.
[0094] The bias variation circuit 200 may vary at least one bias supplied to the photoreceptor 110_2. For example, the power supply voltage bias VB1 supplied to the photoreceptor 110_2, the bias current IBIAS flowing through the constant current source 111 of the logarithmic amplifier circuit, or the gate node voltage bias VB3 of the feedback gain control transistor MFBP may vary. Therefore, even when the intensity of the incident light does not change or changes very little, the bias variation circuit 200 may vary according to the bias voltage supplied to the photoreceptor 110_2. Fig.11 The dynamic vision sensor of the embodiment can also output the event signal ES based on the changed bias, rather than outputting the event signal ES based on the change in the intensity of the incident light. Therefore, even when the intensity of the incident light does not change or changes very little, the normal operation of the application can be guaranteed.
[0095] Fig.12 is a circuit diagram of a photoreceptor 110_3 according to some example embodiments. Fig.11 The photoreceptor 110_2 described is different, Fig.12 The logarithmic amplifier circuit in the photoreceptor 110_3 in FIG. 1 includes a transistor MB4. Detailed description of parts identical or similar to those described above will be omitted to avoid redundancy.
[0096] The drain node of transistor MB4 according to some example embodiments is connected to the gate node of feedback transistor MFBN through node N3. The source node of transistor MB4 is connected to the drain node of logarithmic amplifier transistor MAMP. The amplification degree of the logarithmic amplifier circuit can be finely adjusted by transistor MB4.
[0097] The bias variation circuit 200 may vary at least one bias provided to the photoreceptor 110_3. For example, the power supply voltage bias VB1 provided to the photoreceptor 110_3, the bias current IBIAS flowing through the constant current source 111 of the logarithmic amplifier circuit, the gate node voltage bias VB3 of the feedback gain control transistor MFBP, or the gate node voltage bias VB4 of the transistor MB4 may vary. Therefore, even when the intensity of the incident light does not change or changes very little, the bias variation circuit 200 may vary according to the bias voltage. Fig.12The dynamic vision sensor of the embodiment can also output the event signal ES based on the changed bias, rather than outputting the event signal ES based on the change in the intensity of the incident light. Therefore, even when the intensity of the incident light does not change or changes very little, the normal operation of the application can be guaranteed.
[0098] Fig.13 is a block diagram of a dynamic vision sensor 10_1 according to some example embodiments. Fig.13 The dynamic vision sensor 10_1 of the embodiment may be composed of a plurality of substrates. Detailed description of the parts identical or similar to the above parts will be omitted to avoid redundancy.
[0099] In some example embodiments, the dynamic vision sensor 10_1 may be a stacked dynamic vision sensor including a plurality of vertically stacked substrates electrically connected to each other. Fig.13 The dynamic vision sensor 10_1 described in Figure 3 The embodiments are described under the premise that Figure 4 The following embodiments are illustrative, and the present disclosure does not particularly limit the method of stacking substrates.
[0100] The photoreceptor may be disposed on a first substrate among the plurality of substrates, and the pixel back-end circuit may be disposed on a second substrate electrically connected to the first substrate. The row sampling circuit RSC, the column driver CD, the packetizer, the bias variation controller BVC and the bias generator Bias Gen may be disposed on the second substrate.
[0101] Alternatively or additionally, a voltage buffer may be disposed on the second substrate together with the pixel back-end circuit. The voltage buffer may be a reference Figure 6 The voltage buffer circuit 150 is described.
[0102] Among the stacked substrates, the photoreceptor on the first substrate and the pixel back-end circuit on the second substrate may be electrically connected to each other through an in-pixel interconnector 602. For example, the photoreceptor on the first substrate and the pixel back-end circuit on the second substrate may be electrically connected through an in-pixel Cu-to-Cu (C2C) bond or an in-pixel through-silicon-via (TSV, or through-silicon via).
[0103] The column driver CD on the second substrate may send a selection signal to the photoreceptor on the first substrate through a signal path connected to the C2C junction 601 or the TSV 603 or a signal path connected via the TSV. The column bias change controller BVC on the second substrate may receive a bias change enable / disable signal and may control the bias generator Bias Gen. The bias generator Bias Gen may provide a changed bias to the photoreceptor on the first substrate through a signal path connected via the C2C junction 601 or the TSV 603.
[0104] Therefore, the photoreceptor can be placed on the first substrate while ensuring the light receiving area of the photoreceptor by placing the pixel back-end circuit and the bias varying circuit on the second substrate.
[0105] FIG. 14A to FIG. 14C is a block diagram of a dynamic vision sensor according to some example embodiments. Fig.13 Detailed descriptions of parts that are the same or similar to those described above will be omitted to avoid redundancy.
[0106] In one example embodiment, referring to Fig.14A , the dynamic vision sensor 10_2 may be a stacked dynamic vision sensor including a plurality of vertically stacked substrates “electrically connected to each other in a single chip package”.
[0107] The photoreceptor may be disposed on a first substrate among the plurality of substrates, and at least a portion of the pixel back-end circuit may be disposed on a second substrate electrically connected to the first substrate. The pixel back-end circuit may include a voltage buffer. The voltage buffer may be a reference Figure 6 The voltage buffer circuit 150 described above. The row sampling circuit RSC, the column driver CD and the packetizer may be placed on the third substrate. The bias variation controller BVC and the bias generator Bias Gen may be placed on the second substrate.
[0108] Among the stacked substrates, the photoreceptor on the first substrate and the pixel back-end circuit on the second substrate may be electrically connected to each other through the intra-pixel interconnector 602. For example, the photoreceptor on the first substrate and the voltage buffer of the pixel back-end circuit on the second substrate may be electrically connected through the intra-pixel C2C junction or the intra-pixel TSV. The pixel back-end circuit may be connected through a signal path "connected to the row sampling circuit RSC on the third substrate via the C2C junction 601 or the TSV 603".
[0109] The column driver CD on the third substrate may send a selection signal to the photoreceptor on the first substrate through a signal path connected via the C2C bonding member 601 or a signal path connected via the TSV 603. The column bias change controller BVC on the second substrate may receive a bias change enable / disable signal from the image signal processor ISP on the third substrate and may control the bias generator Bias Gen. The bias generator Bias Gen may provide a changed bias to the photoreceptor on the first substrate through a signal path connected via the C2C bonding member 601 or the TSV 603.
[0110] Therefore, the image signal processor and the dynamic vision sensor 10_2 may be configured as one sensor.
[0111] The image signal processor ISP on the third substrate can reduce noise of the event signal included in the packet, or can correct the packet. For example, the image signal processor ISP can correct the event data packet based on the response time according to the address and illumination value of the pixel where the on event occurs.
[0112] The image signal processor ISP according to some example embodiments may generate a grayscale image by using an event signal based on a changed offset. Fig.18 and Fig.19 A method of generating a grayscale image by using an event signal based on a changed offset by an image signal processor ISP is described.
[0113] In one example embodiment, referring to Fig. 14B , the dynamic vision sensor 10_3 may receive a power supply voltage VDD and a power supply voltage bias from the bias change circuit 200_3. The power supply voltage bias may be one of a power supply voltage bias VB1 and a power supply voltage bias VB2. The bias change circuit 200_3 may adjust the power supply voltage VSUP and generate a power supply voltage VDD and a power supply voltage bias. The bias change circuit 200_3 may provide one of the power supply voltage VDD and the power supply voltage bias based on a bias change enable signal and a bias change disable signal. The bias change circuit 200_3 may be located outside the dynamic vision sensor 10_3. The bias change circuit 200_3 may be located in a chip package different from the chip package in which the dynamic vision sensor 10_3 is located.
[0114] In one example embodiment, referring to Fig. 14C, the dynamic vision sensor 10_3 may receive a power supply voltage VDD from a power supply circuit PS and a power supply voltage bias from a bias change circuit 200_4. The power supply voltage bias may be one of a power supply voltage bias VB1 and a power supply voltage bias VB2. The switch SW may select the power supply voltage VDD or the power supply voltage bias based on a bias change enable signal and a bias change disable signal. The bias change circuit 200_3 and the power supply circuit PS may be located outside the dynamic vision sensor 10_4.
[0115] Fig.15 is a block diagram of an image processing system 1 according to some example embodiments. The image processing system 1 may include a camera 1000 and an application processor 2000. Fig.16 The image processing system 1 may be applied to a portable electronic device (such as a digital camera, a video camera, a mobile phone, a smart phone, a tablet personal computer (PC), a personal digital assistant (PDA), a wearable computer, an augmented reality (AR) device, or an Internet of Things (IoT) device). In addition, the image processing system 1 may be applied to a simultaneous localization and mapping (SLAM) device (such as an autonomous vehicle, an ADAS, a drone, a safety device (such as a safety camera), a mobile robot, etc.).
[0116] The camera 1000 may include a dynamic vision sensor 10, a lens 20, an actuator 30, and an interface (eg, an input / output interface) 40. The actuator 30 may be optionally not included according to an embodiment. Alternatively, the camera 1000 may include an RGB image sensor. The RGB image sensor may be a CMOS image sensor (CIS).
[0117] Under the control of the application processor 2000, the dynamic vision sensor 10 may sense the object 3000 captured through the lens 20. The dynamic vision sensor 10 may generate an event signal based on a change in the intensity of a light signal of the object 3000 incident through the lens 20. The event signal may be packetized and output.
[0118] The dynamic vision sensor 10 may be connected to an external device (such as the application processor 2000) through the input / output interface 40 in an inter-integrated circuit (I2C) or mobile industry processor interface (MIPI) method. The dynamic vision sensor 10 may transmit a packet to the application processor 2000 through the input / output interface 40, or may receive a control signal CS from the application processor 2000.
[0119] The application processor 2000 may receive the event signal through the interface 2100, and the image signal processor 2200 may process the event signal. For example, the image signal processor 2200 may reduce the noise of the event signal, or may correct the packet. For example, the image signal processor 2200 may correct the event data packet based on the response time according to the address and the illuminance value of the pixel where the on event occurs. Alternatively, the application processor 2000 may not include the image signal processor 2200. In this case, the noise processing of the event signal and the correction of the packet may be performed in the core processor 2300.
[0120] The image signal processor 2200 may generate an event image based on the event signal. The event image may be an image obtained by setting simultaneous on events for each pixel value.
[0121] The core processor 2300 operates an application. For example, the application may be used to detect an object from an event signal. According to an embodiment, some operations of the core processor 2300 may be performed by an artificial intelligence (AI) processor 2500 (e.g., a graphics processing unit (GPU), a general purpose graphics processing unit (GPGPU), an AI dedicated processor, a parallel processing processor, etc.) electrically connected to the core processor 2300.
[0122] The dynamic vision sensor 10 according to some example embodiments may include a reference Figures 1 to 14C The photoreceptor and bias variation circuit of the described embodiments.
[0123] Even when the intensity of the light signal of the object 3000 does not change, the dynamic vision sensor 10 according to some example embodiments may output an event signal based on the changed bias provided by the bias change circuit. The operation mode in which the dynamic vision sensor 10 outputs an event signal based on the changed bias provided by the bias change circuit may be referred to as a "bias change mode (or variable mode)". The dynamic vision sensor 10 may operate in the bias change mode in response to a bias change enable signal received from the application processor 2000. The bias change enable signal may be transmitted while being included in the control signal CS.
[0124] The dynamic vision sensor 10 may operate in a normal mode in which an event signal is output based on a change in the intensity of a light signal of the object 3000, and may operate in a bias change mode in response to a bias change enable signal received from the application processor 2000. The dynamic vision sensor 10 may operate in the normal mode again in response to a bias change disable signal received from the application processor 2000.
[0125] The application processor 2000 may control the dynamic vision sensor 10 so that the mode of the dynamic vision sensor 10 is switched between the normal mode and the bias change mode according to an embodiment of the application program. For example, the ADAS of the vehicle may detect an object based on an event signal output by the dynamic vision sensor 10 of the front camera in the normal mode. When the object is not detected based on the event signal output in the normal mode (for example, when the front vehicle is driving at the same speed as the front vehicle), the application processor 2000 may send a bias change enable signal to the camera 1000 to switch the mode of the dynamic vision sensor 10 to the variable mode.
[0126] The image signal processor 2200 or the core processor 2300 of the application processor 2000 may generate a grayscale image based on the event signal of the bias change mode. The image signal processor 2200 or the core processor 2300 may identify the event signal of the bias change mode based on the header of the packet (referred to as "event data"). The header may be an information bit added by a packetizer of the dynamic vision sensor in response to a bias change on signal or a bias change off signal.
[0127] The image signal processor 2200 or the core processor 2300 may detect an object based on a grayscale image. An algorithm for detecting an object based on an event signal in a normal mode may be different from an algorithm for detecting an object based on a grayscale image.
[0128] Fig.17 It is used to describe the Fig.15 Flowchart of an operation of generating a grayscale image by the application processor 2000 of the image processing system 1 of the embodiment and detecting an object based on the grayscale image. The image signal processor 2200 of the application processor 2000 may generate a grayscale image based on event data.
[0129] The core processor 2300 may detect an object by executing an object detection algorithm based on a grayscale image.
[0130] The core processor 2300 may detect an object based on an event signal generated by a dynamic vision sensor in a normal mode.
[0131] In operation S110 , the application processor 2000 may receive event data (eg, an event signal included in the event data) from the dynamic vision sensor 10 of the camera 1000 .
[0132] In operation S120, the application processor 2000 may check the header of the event data received from the dynamic vision sensor (e.g., the type of event signal included in the event data). The application processor 2000 may determine whether the event data is event data generated in the bias variation mode or event data generated in the normal mode based on a predetermined information bit. The application processor 2000 may apply a conventional event signal processing method to the event data generated in the normal mode.
[0133] In operation S130, the application processor 2000 may generate a grayscale image based on an event signal included in event data generated in the bias variation mode.
[0134] Reference Fig.18 , the application processor 2000 generates a reference frame image in response to receiving the first event signal L0. For example, the reference frame image may be a frame image in which all image pixel values are set to the same grayscale level (for example, when a grayscale image of level 256 is generated, each pixel value may be set to 128).
[0135] The application processor 2000 may generate a grayscale image by reflecting the event signals L1 to L6 received after the first event signal L0 to the respective pixel values of the reference frame image. For example, the grayscale image based on the event signal L6 may be an image obtained by accumulating and reflecting the previously received event signals L1 to L5 and the event signal L6 to the respective pixel values of the reference frame image.
[0136] A method in which the application processor 2000 reflects the event signals L1 to L6 to respective pixel values of the reference frame image will be described.
[0137] In the dynamic vision sensor, the event signal L1 generated at time T1 includes relative information about the event signal L0 generated at time T0. For example, when the event signal L1 is an on event signal, the intensity of the incident light contributing to the event signal L1 is greater than the intensity of the incident light contributing to the event signal L0. Therefore, the logarithmic scale voltage of the pixel of the dynamic vision sensor corresponding to the event signal L1 is greater than the logarithmic scale voltage of the pixel of the dynamic vision sensor corresponding to the event signal L0. In addition, when the event signal L3 is an off event signal, the intensity of the incident light contributing to the event signal L3 is less than the intensity of the incident light contributing to the event signal L2. Therefore, the logarithmic scale voltage of the pixel of the dynamic vision sensor corresponding to the event signal L3 is less than the logarithmic scale voltage of the pixel of the dynamic vision sensor corresponding to the event signal L2.
[0138] Therefore, the logarithmically scaled voltage of a pixel of the dynamic vision sensor 10 at a specific time point (t) may be expressed based on the following Equation 1.
[0139] [Equation 1]
[0140] L(t) represents the illuminance of the incident light. "c" represents the logarithmic scale voltage difference between adjacent event signals. L(f) represents the illuminance of the incident light corresponding to the reference frame image. k represents the event signal at time point “k”, and e k = +1 (corresponding to an on event signal) or -1 (corresponding to an off event signal).
[0141] Equation 1 can be summarized as Equation 2 below.
[0142] [Equation 2]
[0143] Equation 2 shows that the illuminance of incident light of a pixel of the dynamic vision sensor 10 at a specific time point (t) may be generated based on an event signal at a previous time point (f).
[0144] Therefore, the application processor 2000 can identify the event type (polarity) of the event signals L1 to L6 received after the first event signal L0, and can change the individual pixel values of the reference frame image according to the event type. In other words, when an on event occurs, the pixel value of the previous frame image can be changed by multiplying the pixel value of the previous frame image by e c To change the pixel value, and when the off event occurs, the pixel value of the previous frame image can be divided by e c to change the pixel value.
[0145] Therefore, each pixel value in the grayscale image is not the precise absolute illuminance value of the incident light contributing to the event signal. However, the relative illuminance difference between the incident lights incident on each pixel of the dynamic vision sensor 10 at the same time can also be reflected.
[0146] Return to Fig.17 , in operation S140, the application processor 2000 may detect an object by applying an object detection algorithm to a grayscale image generated based on the event signal.
[0147] In operation S150, when a predetermined condition is satisfied, the application processor 2000 may transmit a bias change disable signal to the dynamic vision sensor 10. For example, when a certain time has passed since the bias change enable signal was transmitted, the application processor 2000 may transmit the bias change disable signal.
[0148] When the header of the event data checked in operation S120 indicates event data in the normal mode, the application processor 2000 performs operation S160.
[0149] In operation S160, the application processor 2000 may detect an object based on the event signal in the normal mode.
[0150] In operation S170 , when the object is not detected, the application processor 2000 may transmit a bias change enable signal to the dynamic vision sensor 10 .
[0151] Fig.19 It is used to describe the Fig.15 Flowchart of an operation of generating a grayscale image and performing autofocusing based on the grayscale image by the application processor 2000 of the image processing system 1 of the embodiment of the present invention. The image signal processor 2200 of the application processor 2000 may generate a grayscale image based on event data. The core processor 2300 may perform an autofocusing algorithm based on the grayscale image. Fig.17 and Fig.18 Detailed descriptions of those identical or similar parts described will be omitted to avoid redundancy.
[0152] In operation S210 , the application processor 2000 turns on the auto focus mode. The application processor 2200 may transmit a bias change enable signal to the dynamic vision sensor 10 .
[0153] In operation S220, the image signal processor 2200 may perform auto focusing of the camera 1000 based on the contrast of the grayscale image.
[0154] For example, the core processor 2300 may transmit the control signal CS of the actuator 30 to the camera 1000. The core processor 2300 may receive first event data and second event data generated at different positions of the lens 20. The first event data and the second event data are a plurality of pieces of event data generated in the bias variation mode.
[0155] The image signal processor 2200 may generate a first grayscale image and a second grayscale image based on the first event data and the second event data, respectively. Fig.18 The described method generates a grayscale image.
[0156] The image signal processor 2200 may measure a contrast ratio of a specific portion of the image between at least some areas of the first grayscale image and the second grayscale image.
[0157] In operation S230 , when the contrast between the first grayscale image and the second grayscale image is amplified (eg, increased to a maximum or close to a maximum), the image signal processor 2200 may stop the auto focusing. The application processor 2200 may transmit a bias change disable signal to the dynamic vision sensor 10 .
[0158] At the same time, the above description represents detailed embodiments for implementing the present disclosure. Embodiments in which the design is simply changed or easily changed may be included together with the above-mentioned example embodiments. In addition, technologies that are easily changed and implemented by using the above-mentioned embodiments may be included here. Although the inventive concept has been described with reference to the above-mentioned embodiments, it is obvious to those of ordinary skill in the art that various changes and modifications may be made thereto without departing from the spirit and scope of the example embodiments as set forth in the appended claims.
[0159] Even when the intensity of incident light does not change, the image sensor according to some example embodiments may generate an event signal.
[0160] Any element and / or functional block disclosed above may include a processing circuit (such as hardware including a logic circuit, a hardware / software combination (such as a processor that executes software), or a combination thereof) or be implemented in a processing circuit (such as hardware including a logic circuit, a hardware / software combination (such as a processor that executes software), or a combination thereof). For example, the processing circuit may more specifically include, but is not limited to, a central processing unit (CPU), an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), a system on a chip (SoC), a programmable logic unit, a microprocessor, an application specific integrated circuit (ASIC), etc. The processing circuit may include an electrical component (such as at least one of a transistor, a resistor, a capacitor, etc.). The processing circuit may include an electronic component (such as a logic gate including at least one of an AND gate, an OR gate, a NAND gate, a NOT gate, etc.).
[0161] The processor according to some example embodiments may generate a grayscale image based on the event signal.
[0162] Although the present disclosure has been described with reference to the embodiments of the present disclosure, it will be apparent to those skilled in the art that various changes and modifications may be made thereto without departing from the spirit and scope of the present disclosure as set forth in the appended claims. In addition, the example embodiments are not necessarily mutually exclusive. For example, some example embodiments may include one or more features described with reference to one or more figures, and may also include one or more other features described with reference to one or more other figures.
Claims
1. A dynamic vision sensor, comprising: a photodiode configured to generate a first photodiode current in response to a change in light due to movement of the object; a first transistor; a first node connected to the photodiode and the first transistor; a second node connected to the first transistor and configured to receive the first voltage and to receive the second voltage; a second transistor connected to the first node and ground; a third node connected to the first transistor and the second transistor; as well as A current source, connected to the third node, The first voltage has a first level, and the second voltage has a second level different from the first level.
2. The dynamic vision sensor according to claim 1, wherein: The third node is connected to the gate of the first transistor, and The first node is connected to the gate of the second transistor.
3. The dynamic vision sensor according to claim 1, wherein: The photodiode is configured to generate a second photodiode current in response to the first voltage.
4. The dynamic vision sensor according to claim 1, wherein: The photodiode is configured to generate a third photodiode current in response to the second voltage.
5. The dynamic vision sensor according to claim 1, wherein: The second node is configured to receive a third voltage, and The third voltage has a third level different from at least one of the first level and the second level.
6. The dynamic vision sensor according to claim 1, wherein: The second node is configured to alternately receive the first voltage and the second voltage.
7. The dynamic vision sensor according to claim 1, wherein: The second node is configured to receive a first voltage from a bias varying circuit located in a first chip package that is different from a second chip package where the photodiode is located.
8. The dynamic vision sensor according to claim 1, wherein: The second node is configured to: receive a first voltage from the bias varying circuit, and The bias variation circuit and the photodiode are located in a single chip package.
9. The dynamic vision sensor according to claim 1, wherein: The current source is configured to receive a fourth voltage from the bias varying circuit.
10. The dynamic vision sensor according to claim 9, wherein: The current source is configured to receive a fifth voltage different from the fourth voltage from the bias varying circuit.
11. The dynamic vision sensor according to any one of claims 1 to 10, wherein: The bias varying circuit is configured to supply a first voltage and a second voltage to a second node based on a bias varying enable signal.
12. The dynamic vision sensor according to any one of claims 1 to 10, wherein: The bias varying circuit is configured not to supply the first voltage and the second voltage to the second node based on the bias varying disable signal.
13. A dynamic vision sensor, comprising: a photodiode configured to generate a first photodiode current in response to a change in light due to movement of the object; a first transistor; a first node connected to the photodiode and the first transistor; a second node connected to the first transistor; a second transistor connected to the first node and ground; a third node connected to the first transistor and the second transistor; as well as a current source connected to the third node and configured to receive the first voltage and the second voltage, The first voltage has a first level, and the second voltage has a second level different from the first level.
14. The dynamic vision sensor according to claim 13, wherein: The third node is connected to the gate of the first transistor, and The first node is connected to the gate of the second transistor.
15. The dynamic vision sensor according to claim 13, wherein: The photodiode is configured to generate a second photodiode current in response to the first voltage.
16. The dynamic vision sensor according to claim 13, wherein: The photodiode is configured to generate a third photodiode current in response to the second voltage.
17. The dynamic vision sensor according to claim 13, wherein: The current source is configured to receive a third voltage having a third level different from at least one of the first level and the second level.
18. The dynamic vision sensor according to claim 13, wherein: The current source is configured to alternately receive a first voltage and a second voltage.
19. The dynamic vision sensor according to any one of claims 13 to 18, wherein: The third node is configured to receive a first voltage from a bias varying circuit located in a first chip package that is different from a second chip package where the photodiode is located.
20. The dynamic vision sensor according to any one of claims 13 to 18, wherein: The third node is configured to: receive a first voltage from the bias varying circuit, and The bias variation circuit and the photodiode are located in a single chip package.
Citation Information
Patent Citations
Method for manufacturing fermented food having an effect on liver health and fermented food produced by the method
KR1020230091355A