WO2021038751A1 - センサシステム、画像処理装置、画像処理方法およびプログラム - Google Patents
センサシステム、画像処理装置、画像処理方法およびプログラム Download PDFInfo
- Publication number
- WO2021038751A1 WO2021038751A1 PCT/JP2019/033708 JP2019033708W WO2021038751A1 WO 2021038751 A1 WO2021038751 A1 WO 2021038751A1 JP 2019033708 W JP2019033708 W JP 2019033708W WO 2021038751 A1 WO2021038751 A1 WO 2021038751A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- sensor
- event
- event signal
- sensitivity
- gradation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N25/00—Circuitry of solid-state image sensors [SSIS]; Control thereof
- H04N25/70—SSIS architectures; Circuits associated therewith
- H04N25/76—Addressed sensors, e.g. MOS or CMOS sensors
- H04N25/77—Pixel circuitry, e.g. memories, A/D converters, pixel amplifiers, shared circuits or shared components
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N25/00—Circuitry of solid-state image sensors [SSIS]; Control thereof
- H04N25/50—Control of the SSIS exposure
- H04N25/57—Control of the dynamic range
- H04N25/58—Control of the dynamic range involving two or more exposures
- H04N25/581—Control of the dynamic range involving two or more exposures acquired simultaneously
- H04N25/585—Control of the dynamic range involving two or more exposures acquired simultaneously with pixels having different sensitivities within the sensor, e.g. fast or slow pixels or pixels having different sizes
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N25/00—Circuitry of solid-state image sensors [SSIS]; Control thereof
- H04N25/47—Image sensors with pixel address output; Event-driven image sensors; Selection of pixels to be read out based on image data
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N25/00—Circuitry of solid-state image sensors [SSIS]; Control thereof
- H04N25/70—SSIS architectures; Circuits associated therewith
- H04N25/79—Arrangements of circuitry being divided between different or multiple substrates, chips or circuit boards, e.g. stacked image sensors
Definitions
- the present invention relates to a sensor system, an image processing device, an image processing method and a program.
- Image sensors such as CCD and CMOS are synchronous solid-state image sensors that image image data (frames) in synchronization with synchronization signals such as vertical synchronization signals.
- image data is acquired only every synchronization signal cycle (for example, 1/60 second), so it is difficult to handle high-speed processing using image data in a moving body, for example. There are cases. Therefore, for example, in Non-Patent Document 1, an asynchronous solid-state image sensor provided with an address event representation (AER: Address Event Representation) circuit for detecting an address event has been proposed.
- AER Address Event Representation
- an address event occurs when the amount of light of a pixel fluctuates at a certain pixel address and the amount of fluctuation exceeds a threshold value.
- the address event includes an on-event that occurs when the light amount of the pixel fluctuates and exceeds a predetermined upper limit, and an off-event that occurs when the light amount falls below a predetermined lower limit.
- the format of image data that expresses the presence or absence of on-events and off-events for each pixel with 2-bit data is called an AER format.
- a technique using an asynchronous solid-state image sensor is also described in, for example, Patent Document 1.
- asynchronous solid-state image sensor (hereinafter, also referred to as an event-driven sensor) as described above can detect the movement of the subject at high speed
- a synchronous solid-state image sensor capable of detecting gradation is arranged together with an event-driven sensor, and when the movement of the subject is detected by the event-driven sensor, the synchronous solid-state image sensor is exposed.
- the acquisition cycle of the image data including the gradation is restricted by the cycle of the synchronization signal of the synchronous solid-state image sensor, so that the high speed of the event-driven sensor is increased. It will be damaged.
- an object of the present invention is to provide a sensor system, an image processing device, an image processing method, and a program that enable high-speed detection of the gradation of a subject using an event-driven sensor.
- the light amount at the first sensor that detects the fluctuation of the light amount at the first pixel address with the first sensitivity, and the light amount at the second pixel address adjacent to or overlapping the first pixel address.
- a sensor array containing a second sensor that detects fluctuations with a second sensitivity lower than the first sensitivity, and a second when the first sensor generates a first event signal for a luminance fluctuation event.
- a sensor system including a gradation determination unit that determines the gradation of a subject that has generated a brightness fluctuation event depending on whether or not the sensor has generated a second event signal for the brightness fluctuation event. ..
- the first event signal generated by the first sensor that detects the fluctuation of the amount of light at the first pixel address with the first sensitivity for the luminance fluctuation event is the first.
- An image processing device including an event signal associating unit to be associated and a gradation determination unit for determining the gradation of a subject that has generated a luminance fluctuation event according to the presence or absence of a second event signal associated with the first event signal. Provided.
- the step of generating the first event signal for the luminance fluctuation event by the first sensor that detects the fluctuation of the amount of light at the first pixel address with the first sensitivity is a second event with respect to a brightness fluctuation event.
- the step of generating a signal or not generating a second event signal the step of associating the first event signal with the second event signal, and the presence or absence of the second event signal associated with the first event signal.
- An image processing method including a step of determining the gradation of the subject in which the brightness fluctuation event is generated is provided.
- the first event signal generated by the first sensor that detects the fluctuation of the amount of light at the first pixel address with the first sensitivity for the luminance fluctuation event is the first.
- a program is provided for the computer to realize a function of associating the image and a function of determining the gradation of the subject in which the brightness fluctuation event is generated according to the presence or absence of the second event signal associated with the first event signal.
- FIG. 1 is a diagram showing a schematic configuration of a system according to a first embodiment of the present invention.
- the sensor system 10A includes a sensor module 100 and an image processing device 200A.
- the sensor module 100 includes a sensor array including a first sensor 111 and a second sensor 112 arranged for each pixel, and a signal processing circuit 120 (event signal processing unit).
- the image processor 200A is implemented by, for example, a computer having a communication interface, a processor, and a memory, and the time difference acquisition realized by the processor operating according to a program stored in the memory or received through the communication interface. Includes functional parts of unit 210A and gradation calculation unit 220A.
- the image processing device 200A may further include functional parts of the image generation unit 230 and the delay time calculation unit 240. Hereinafter, each part will be further described.
- Both the first sensor 111 and the second sensor 112 are event-driven sensors (EDS: Event Driven Sensor), and an address event occurs when the fluctuation amount of the light amount exceeds the threshold value at each pixel address. Outputs an event signal indicating.
- the sensor array includes a first sensor 111 and a second sensor 112 arranged in a plane in a predetermined pattern. The first sensor 111 detects the fluctuation of the light amount at the first pixel address with the first sensitivity, and the second sensor 112 detects the fluctuation of the light amount at the second pixel address adjacent to the first pixel address. Detect with a second sensitivity that is lower than the first sensitivity.
- the sensitivity of the second sensor 112 is p times (1> p> 0) of the sensitivity of the first sensor 111 (first sensitivity).
- a first sensor 111 and a second sensor 112 are provided with, for example, a filter 115 (for example, a gray filter or a diaphragm) that is superimposed on the sensor array and reduces the amount of light incident on the second sensor 112. Can be achieved by.
- EDS having the same configuration can be used for the first sensor 111 and the second sensor 112.
- the filter 115 blocks (1-p) times the amount of light
- the second sensitivity is p times the first sensitivity.
- the sensitivity of each sensor may be adjusted by making the bias current different between the first sensor 111 and the second sensor 112.
- the signal processing circuit 120 includes a memory and a processor, and the first event signal generated by the first sensor 111 and the second generated by the second sensor 112 when the processor operates according to a program stored in the memory. Process the event signal of. Specifically, the signal processing circuit 120 generates time stamps of the first event signal and the second event signal, respectively.
- the time stamp is an example of information indicating the difference in time when the first sensor 111 and the second sensor 112 generate event signals for the luminance fluctuation event, respectively.
- the first sensor 111 and the second sensor 112 detect a fluctuation amount of the amount of light when a brightness fluctuation event (hereinafter, also simply referred to as an event) such as a movement of a subject or a change of a light source occurs.
- An event signal is generated when the threshold value is exceeded.
- the sensitivity for detecting fluctuations in the amount of light differs between the first sensor 111 and the second sensor 112. Therefore, even when both the first sensor 111 and the second sensor 112 generate an event signal for the same event, the amount of fluctuation in the amount of light detected by each sensor is different.
- the first sensor 111 and the second sensor utilize the characteristic of the sensor that the delay time from the event to the generation of the event signal is different when the fluctuation amount of the light amount is different.
- the gradation of the subject is calculated from the difference in time when the event signals are generated for each of the same brightness fluctuation events of 112.
- FIGS. 2A, 2B and 3 are diagrams for explaining the principle of gradation calculation in the embodiment of the present invention.
- the luminance variation of an event is relatively large.
- a delay time d 1 occurs from the occurrence of the event until the EDS generates the event signal.
- an event is generated by the movement of the dark (low gradation) subject obj 2 , the amount of brightness fluctuation of the event is relatively small.
- the delay time d 2 that occurs from the occurrence of the event to the generation of the event signal by EDS is longer than the delay time d 1 in the case of the subject obj 1 (d 1 ⁇ d 2 ). That is, the delay time for the EDS to generate the event signal becomes shorter as the gradation of the subject is higher, and becomes longer as the gradation of the subject is lower.
- the first delay time of the event signal d 1 and the difference between the delay time d 2 of the second event signal generated by the second sensor 112 generated by the first sensor 111 (d 2 - d 1 ) corresponds to 50% of the gradation of the subject. Therefore, if the relationship between the gradation of the subject and the delay time of EDS is measured in advance, the gradation of the subject can be calculated from the difference in delay time, that is, the difference in the time when each sensor generates an event signal. it can.
- the gradation of the subject is measured in advance by measuring the relationship between the gradation of the subject that causes the event and the delay time of EDS, so that the gradation of the subject is a function g (d) of the delay time d. Identify as.
- the time difference acquisition unit 210A causes the first sensor and the second sensor to respond to the same luminance fluctuation event based on the time stamp generated by the signal processing circuit 120.
- Information indicating the difference between the times when the event signals were generated is acquired.
- the signal processing circuit 120 has a first event signal and a second event signal apart from the time stamps of the first event signal and the second event signal. Information indicating the time difference between the two event signals may be directly acquired.
- the gradation calculation unit 220A calculates the gradation of the subject in which the luminance fluctuation event has occurred, based on the information indicating the time difference acquired by the time difference acquisition unit 210A. As described above with reference to FIGS. 2A, 2B and 3, if the relationship between the gradation of the subject that generates the event and the delay time of EDS is measured in advance, the time when the event signal is generated The gradation of the subject can be calculated from the difference between the two.
- the time difference-gradation table 221 is stored in the memory of the image processing device 200A, and the gradation calculation unit 220A calculates the gradation of the subject by referring to the time difference-gradation table 221.
- the gradation of the subject calculated by the gradation calculation unit 220A by the above processing may be used, for example, in the image generation unit 230 to generate an image of the subject by using the event signal.
- the first event signal and the second event signal may be synchronized by a process as described later.
- the gradation of the subject calculated by the gradation calculation unit 220A may be used to select a process for the subject recognized from the event signal.
- the functional portion that executes the above processing may be mounted inside the image processing device 200A, or may be mounted in an external device that receives information indicating the gradation of the subject from the image processing device 200A.
- the delay time calculated by the delay time calculation unit 240 as described above may be used, for example, to specify the true occurrence time of the event excluding the delay time.
- the true occurrence time of the event detected by the first sensor 111 can be specified by subtracting the delay time d 1 from the time stamp of the first event signal generated by the first sensor 111. it can.
- the image processing device 200A includes both the gradation calculation unit 220A and the delay time calculation unit 240, but in other embodiments, the image processing device is either the gradation calculation unit 220A or the delay time calculation unit 240. May include only.
- FIGS. 4A to 4D are diagrams for explaining the simultaneous processing in the first embodiment of the present invention.
- the delay time from the occurrence of the luminance fluctuation event to the generation of the event signal differs between the first sensor 111 and the second sensor 112.
- the gradation of the subject is calculated by using the difference in the delay time, but on the other hand, the first event signal generated by the first sensor 111 and the second sensor 112 are generated.
- a difference in time stamp corresponding to the difference in delay time occurs between the signal and the second event signal.
- the time stamps of the second event signals generated by the second sensor 112A are the average values of the time stamps of the event signals generated by the adjacent first sensors 111A to 111H.
- the second event signal is synchronized with the first event signal.
- 4A to 4D show examples of cases where the first sensors 111 that generate event signals adjacent to the second sensor 112A are 2, 4, and 8, respectively.
- the first pixel address and the first pixel address where the first sensor 111 is arranged and the first sensor 111 are generated.
- the movement of the subject can be identified based on the events detected in all the pixels of the sensor array including both of the second pixel addresses where the two sensors 112 are located (ie, without reducing the resolution).
- the simultaneous processing as described above does not necessarily have to be executed.
- the first pixel in which the first sensor 111 is arranged is based on the first event signal generated by the first sensor 111.
- the movement of the subject may be specified based only on the event detected by the address.
- the resolution is reduced by the amount that the event detected at the second pixel address is not used, but the reduction in resolution can be compensated by a known interpolation method such as linear interpolation.
- only the event signal acquired at the first pixel address may be used for identifying the movement of the subject without performing the simultaneous processing or the interpolation processing.
- the gradation is calculated with the minimum necessary resolution while calculating the gradation.
- the resolution for identifying the movement of the subject may be maintained.
- FIG. 5 is a flowchart showing an example of processing in the first embodiment of the present invention.
- the first sensor 111 of the sensor module 100 generates a first event signal for the luminance fluctuation event (step S101).
- the second sensor 112 generates a second event signal for the same luminance variation event (step S102).
- the time difference between the first event signal and the second event signal is such that the detected luminance fluctuation amount is the first sensor 111 and the second sensor 112 having lower sensitivity. It is caused by the difference between.
- the time difference acquisition unit 210A provides information indicating the difference in time when the first sensor 111 and the second sensor 112 generate the first event signal and the second event signal, respectively. Acquire (step S103). Specifically, in the time difference acquisition unit 210A, the first sensor and the second sensor have the same brightness based on the time stamps of the first event signal and the second event signal generated by the signal processing circuit 120. Information indicating the difference between the times when the event signals are generated (steps S101 and S102 above) for each variable event is acquired.
- the gradation calculation unit 220A calculates the gradation of the subject in which the luminance fluctuation event is generated based on the information indicating the time difference acquired by the time difference acquisition unit 210A (step S104). As described with reference to FIG. 1, at this time, the gradation calculation unit 220A may refer to the time difference-gradation table 221 stored in the memory of the image processing device 200A. In the illustrated example, the image generation unit 230 generates an image of the subject using the calculated gradation and the event signal (step S105), but the gradation calculated as described above is combined with the image generation. Alternatively, it may be used for other processing instead of image generation.
- FIG. 6 is a diagram showing another example of the sensor arrangement in the first embodiment of the present invention.
- the sensor array of the sensor module 100 includes a first sensor 111, a second sensor 112, a third sensor 113 and a fourth sensor 114 arranged in a plane in a predetermined pattern.
- the third sensor 113 is adjacent to at least one of the first and second pixel addresses. The fluctuation of the amount of light at the third pixel address is detected with a third sensitivity lower than the second sensitivity.
- the fourth sensor 114 detects fluctuations in the amount of light at a fourth pixel address adjacent to at least one of the first to third pixel addresses with a fourth sensitivity lower than the third sensitivity.
- the signal processing circuit 120 (not shown) generates time stamps of event signals generated by the first to fourth sensors 111 to 114, respectively.
- Such first to fourth sensors 111 to 114 are, for example, a filter 116 (for example, a gray filter or an aperture) that is superposed on the sensor array and reduces the amount of light incident on the second to fourth sensors 112 to 114.
- the filter 116 blocks 25% of the amount of light incident on the second sensor 112, blocks 50% of the amount of light incident on the third sensor 113, and is incident on the fourth sensor 114. Blocks 75% of the amount of light.
- the brightness change of the event detected by the second sensor 112 is 75% and 50% of the brightness change of the event detected by the first sensor 111, respectively. % And 25%, and the difference between the delay times d 1 to d 4 of the first to fourth event signals generated by the first to fourth sensors 111 to 114 (d 4- d 1 ), (d). 3- d 1 ) and (d 2- d 1 ) correspond to 75%, 50%, and 25% of the gradation of the subject, respectively.
- the delay time d 1 satisfying the condition can be accurately searched.
- the gradation calculation result may be stabilized by averaging the gradations calculated from the time difference from the signal.
- sensors having different sensitivities for detecting fluctuations in the amount of light are arranged at adjacent pixel addresses, and the gradation of the subject is determined from the time difference of the event signal generated by each sensor. calculate.
- the gradation of the subject can be detected without impairing the high speed of the event-driven sensor (EDS).
- EDS event-driven sensor
- FIG. 7 is a diagram showing a schematic configuration of a system according to a second embodiment of the present invention.
- the sensor system 10B includes a sensor module 300 and an image processing device 200A.
- the sensor module 300 includes a sensor array including a stacked sensor 310 arranged for each pixel, and a signal processing circuit 320 (event signal processing unit). Since the configuration of the image processing device 200A is the same as that of the first embodiment described above, duplicate description will be omitted. Hereinafter, each part will be further described.
- the transmission type first light receiving layer 311 constituting the first sensor and the second light receiving layer 312 constituting the second sensor are laminated, and the first light receiving layer 311 is formed. It is arranged on the incident side of light, that is, on the side closer to the subject. The light from the subject passes through the first light receiving layer 311 and is also incident on the second light receiving layer 312.
- the first light receiving layer 311 detects the fluctuation of the light amount at the pixel address where the sensor 310 is arranged with the first sensitivity
- the second light receiving layer 312 detects the fluctuation of the light amount at the same pixel address with the first sensitivity. Detect with a low second sensitivity.
- a first pixel address in which a first sensor composed of a first light receiving layer 311 is arranged and a second sensor in which a second sensor composed of a second light receiving layer 312 is arranged are arranged. It overlaps with the pixel address of.
- the sensitivity of the second light receiving layer 312 is p times (1>p> 0) of the sensitivity of the first light receiving layer 311 (first sensitivity).
- the quantum efficiency of the first light receiving layer 311 is p 1 (1> p 1 > 0)
- the quantum efficiency of the second light receiving layer 312 is p 2 (((1> p 1> 0)). This can be achieved by 1-p 1 )> p 2> 0).
- the quantum efficiency means an index indicating the probability of detection per photon.
- the EDS 310 may have a laminated structure of more than two layers.
- the signal processing circuit 320 includes a memory and a processor, and the first event signal generated by the first sensor configured by the first light receiving layer 311 by operating the processor according to the program stored in the memory, and the first event signal.
- the second event signal generated by the second sensor composed of the light receiving layer 312 of 2 is processed.
- the signal processing circuit 320 generates time stamps for each of the first event signal and the second event signal.
- the time stamp is the difference between the times when the first sensor composed of the first light receiving layer 311 and the second sensor composed of the second light receiving layer 312 generate event signals for the luminance fluctuation event. This is an example of information indicating.
- the image generation unit 230 may specify the movement of the subject by using only the first event signal.
- FIG. 8 is a diagram showing another example of the sensor arrangement in the second embodiment of the present invention.
- the first light receiving layer 311, the second light receiving layer 312, the third light receiving layer 313, and the fourth light receiving layer 314 are laminated in this order from the light incident side. ..
- the first to third light receiving layers 311 to 313 are transmission type, and the light from the subject passes through the first light receiving layer 311 and the second light receiving layer 312 and the third light receiving layer 313 to receive the fourth light. It is incident up to layer 314.
- the first light receiving layer 311 and the second light receiving layer 312 similar to the example described with reference to FIG.
- the third light receiving layer 313 has the amount of light at the pixel address where the sensor 310 is arranged. Fluctuations are detected with a third sensitivity lower than the second sensitivity to form a third sensor.
- the fourth light receiving layer 314 detects fluctuations in the amount of light at the same pixel address with a fourth sensitivity lower than the third sensitivity, and constitutes a fourth sensor.
- the signal processing circuit 120 generates time stamps of event signals generated by sensors composed of the first to fourth light receiving layers 311 to 314, respectively.
- the sensitivity of the second light receiving layer 312 (second sensitivity), the sensitivity of the third light receiving layer 313 (third sensitivity), and the sensitivity of the fourth light receiving layer 314 (fourth).
- sensitivity is, p 2 times the respective sensitivities of the first light-receiving layer 311 (first sensitivity), p 3 times, and p 4 times (1> p 2> p 3 > p 4> 0).
- p 2 0.5
- p 3 0.25
- the quantum efficiency of the first light receiving layer 311 is 40%
- the quantum efficiency of the second light receiving layer 312 is 20%
- the quantum efficiency of the third light receiving layer 313 is 10%
- the change in the brightness of the event detected by the second light receiving layer 312, the third light receiving layer 313, and the fourth light receiving layer 314 is the change in the brightness of the event detected by the first light receiving layer 311, respectively. It becomes 50%, 25% and 12.5%, and the difference between the delay times d 1 to d 4 of the first to fourth event signals generated in the first to fourth light receiving layers 311 to 314, respectively (d 4).
- -D 1 ), (d 3- d 1 ) and (d 2- d 1 ) correspond to 50%, 25% and 12.5% of the gradation of the subject, respectively.
- the gradation of the subject is determined from the time difference of the event signal generated by each light receiving layer. Is calculated.
- the gradation of the subject can be detected without impairing the high speed of the event-driven sensor (EDS).
- EDS event-driven sensor
- FIG. 9 is a diagram showing a schematic configuration of a system according to a third embodiment of the present invention.
- the sensor system 10C includes a sensor module 300 including a sensor 310 having a four-layer laminated structure, and an image processing device 200C.
- the sensor module 300 includes a sensor array including a stacked sensor 310 and a signal processing circuit 320 as in the second embodiment described above.
- the image processor 200C is implemented by, for example, a computer having a communication interface, a processor, and a memory, and the event signal association realized by the processor operating according to a program stored in the memory or received via the communication interface. Includes functional parts of unit 210C and gradation determination unit 220C.
- the image processing device 200C may further include a functional portion of the image generation unit 230.
- each part of the image processing apparatus 200C will be further described.
- the event signal association unit 210C associates the second to fourth event signals with the first event signal input from the signal processing circuit 320.
- the first event signal is generated by the first sensor configured by the first light receiving layer 311 for the luminance fluctuation event.
- the second to fourth sensors composed of the second to fourth light receiving layers 312 to 314 have the same luminance fluctuation as the first sensor. It was generated for the event.
- the event signal association unit 210C executes the association of event signals based on, for example, the order of each event signal and the interval of time stamps.
- the second event signal is generated after the first event signal, and the time stamp interval between the first event signal and the second event signal is within a predetermined range.
- the second event signal is associated with the first event signal. If there is no such second event signal, the event signal association unit 210C does not associate the second event signal with the first event signal.
- the third event signal is generated after the second event signal, and the time stamp interval between the second event signal and the third event signal is within a predetermined range. In some cases, the third event signal is associated with the first event signal. When there is no such third event signal, and when the second event signal is not associated with the first event signal, the event signal associating unit 210C associates the third event signal with the first event signal. Absent.
- the gradation determination unit 220C determines the gradation of the subject according to the result of the association of the event signals in the event signal association unit 210C.
- the first to fourth light receiving layers 311 to 314 of the stacked sensor 310 are configured so that the sensitivity to fluctuations in the amount of light gradually decreases, so that the brightness fluctuates.
- the fluctuation amount of the light amount does not exceed the threshold value in the light receiving layer having relatively low sensitivity, and the sensor composed of the light receiving layer does not generate an event signal.
- the second to fourth light receiving layers 312 to 314 also generate the event signal, but the brightness change is small.
- the fourth light receiving layer 314, which has the lowest sensitivity, does not generate an event signal, and when the brightness change becomes smaller, the third light receiving layer 313 also does not generate an event signal, and the smallest detectable brightness change.
- the second light receiving layer 312 also does not generate an event signal, and only the first light receiving layer 311 generates an event signal.
- the gradation determination unit 220C determines the gradation of the subject according to the relationship between the gradation of the subject and the generation of each event signal as described above. Specifically, when all of the second to fourth event signals associated with the first event signal by the event signal association unit 210C are present, the gradation determination unit 220C has the highest level of gradation of the subject. Is determined to be. On the other hand, when the second and third event signals associated with the first event signal are present but the fourth event signal is not present, the gradation determination unit 220C has the gradation of the subject at the second level. Judge that there is.
- the gradation determination unit 220C determines that the gradation of the subject is the third level. When there is no event signal associated with the first event signal, the gradation determination unit 220C determines that the gradation of the subject is the lowest level.
- the gradation of the subject determined in this way is used, for example, for generating an image of the subject by the image generation unit 230 and selecting processing for the subject recognized from the event signal, as in the first embodiment. Be done.
- FIG. 10 is a flowchart showing an example of processing according to the third embodiment of the present invention.
- FIG. 10 describes the determination using the first event signal and the second event signal, but the same applies to the determination using the third event signal and the fourth event signal.
- the first light receiving layer 311 constituting the first sensor in the sensor module 300 generates a first event signal for the luminance fluctuation event (step S201).
- the second light receiving layer 312 constituting the second sensor generates a second event signal for the same luminance fluctuation event (step S202). If the fluctuation amount of the light amount does not exceed the threshold value in the second light receiving layer 312 having a relatively low sensitivity, the second event signal is not generated.
- the event signal association unit 210C determines whether or not the second event signal is associated with the first event signal (step S203). Specifically, the event signal association unit 210C determines whether or not the association is possible based on the order of the first event signal and the second event signal indicated by the time stamps and the time stamp interval. When the second event signal is associated with the first event signal, the gradation determination unit 220C determines that the gradation of the subject is at a relatively high level (step S204). On the other hand, when the second event signal is not associated with the first event signal, the gradation determination unit 220C determines that the gradation of the subject is at a relatively low level (step S205).
- the senor 310 has a four-layer laminated structure including the first to fourth light receiving layers 311 to 314, but in the other example, the sensor 310 is similar to the example shown in FIG. It may have a laminated structure of layers, or may have a laminated structure of three layers or more than four layers.
- the sensor module may include a sensor array in which sensors of different sensitivities are arranged in a plane in a predetermined pattern, as described above with reference to FIGS. 1 and 6.
- the gradation of the subject is determined depending on which light receiving layer generates the event signal. judge.
- the gradation of the subject can be detected without impairing the high speed of the event-driven sensor (EDS).
- the gradation of the subject is detected only at the same stage as the number of sensitivities of the light receiving layer, but it is convenient because it is not necessary to measure the relationship between the gradation of the subject and the delay time of EDS in advance.
- the gradation of the subject can be stably specified by a simple determination.
- a game controller for example, a game controller, a smartphone, and various moving objects (automobiles, electric vehicles, hybrid electric vehicles, motorcycles, bicycles, personal mobility, airplanes, drones, ships, robots, etc.) are used to provide information on the surrounding environment. It can be used to acquire, estimate the self-position from the position of the surrounding subject, detect the flying subject, and take an avoidance action.
- the gradation of the subject can be useful for identifying the subject or identifying the subject in the above-mentioned applications, for example.
- 10A, 10B, 10C ... sensor system 100 ... sensor module, 111 ... first sensor, 112 ... second sensor, 113 ... third sensor, 114 ... fourth sensor, 115, 116 ... filter, 120 ... Signal processing circuit, 200A, 200C ... Image processing device, 210A ... Time difference acquisition unit, 210C ... Event signal association unit, 220A ... Gradation calculation unit, 220C ... Gradation determination unit, 230 ... Image generation unit, 240 ... Delay time Calculation unit, 300 ... sensor module, 310 ... sensor, 311 ... first light receiving layer, 312 ... second light receiving layer, 313 ... third light receiving layer, 314 ... fourth light receiving layer, 320 ... signal processing circuit.
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Transforming Light Signals Into Electric Signals (AREA)
- Studio Devices (AREA)
Abstract
Description
図1は、本発明の第1の実施形態に係るシステムの概略的な構成を示す図である。図示されているように、センサシステム10Aは、センサモジュール100と、画像処理装置200Aとを含む。センサモジュール100は、画素ごとに配置された第1のセンサ111および第2のセンサ112を含むセンサアレイと、信号処理回路120(イベント信号処理部)とを含む。画像処理装置200Aは、例えば通信インターフェース、プロセッサ、およびメモリを有するコンピュータによって実装され、プロセッサがメモリに格納された、または通信インターフェースを介して受信されたプログラムに従って動作することによって実現される時刻差分取得部210Aおよび階調算出部220Aの機能部分を含む。画像処理装置200Aは、さらに、画像生成部230および遅延時間算出部240の機能部分を含んでもよい。以下、各部についてさらに説明する。
g(d1+dD)=0.5×g(d1) ・・・(式1)
となるような第1のセンサ111の遅延時間d1と、第1のセンサ111で検出されている被写体の階調g(d1)と算出することができる。
図7は、本発明の第2の実施形態に係るシステムの概略的な構成を示す図である。図示された例において、センサシステム10Bは、センサモジュール300と、画像処理装置200Aとを含む。センサモジュール300は、画素ごとに配置された積層型のセンサ310を含むセンサアレイと、信号処理回路320(イベント信号処理部)とを含む。なお、画像処理装置200Aの構成は、上記の第1の実施形態と同様であるため重複した説明は省略する。以下、各部についてさらに説明する。
図9は、本発明の第3の実施形態に係るシステムの概略的な構成を示す図である。図9に示された例は、上記で図8を参照して説明した例の変形例として説明される。つまり、本実施形態において、センサシステム10Cは、4層の積層構造を有するセンサ310を含むセンサモジュール300と、画像処理装置200Cとを含む。センサモジュール300は、上記の第2の実施形態と同様に積層型のセンサ310を含むセンサアレイと、信号処理回路320とを含む。画像処理装置200Cは、例えば通信インターフェース、プロセッサ、およびメモリを有するコンピュータによって実装され、プロセッサがメモリに格納された、または通信インターフェースを介して受信されたプログラムに従って動作することによって実現されるイベント信号関連付け部210Cおよび階調判定部220Cの機能部分を含む。画像処理装置200Cは、さらに、画像生成部230の機能部分を含んでもよい。以下、画像処理装置200Cの各部についてさらに説明する。
Claims (7)
- 第1の画素アドレスにおける光量の変動を第1の感度で検出する第1のセンサ、および前記第1の画素アドレスに隣接または重複する第2の画素アドレスにおける光量の変動を前記第1の感度よりも低い第2の感度で検出する第2のセンサを含むセンサアレイと、
前記第1のセンサが輝度変動イベントに対して第1のイベント信号を生成したときに、前記第2のセンサが前記輝度変動イベントに対して第2のイベント信号を生成したか否かに応じて前記輝度変動イベントを発生させた被写体の階調を判定する階調判定部と
を備えるセンサシステム。 - 前記センサアレイは、前記第1のセンサを構成する透過型の第1の受光層に前記第2のセンサを構成する第2の受光層が積層された積層型のセンサを含み、
前記第2の画素アドレスは、前記第1の画素アドレスに重複する、請求項1に記載のセンサシステム。 - 前記第2の受光層は、透過型であり、
前記積層型のセンサでは、前記第1の画素アドレスにおける光量の変動を前記第2の感度よりも低い第3の感度で検出する第3のセンサを構成する第3の受光層がさらに積層され、
前記階調判定部は、前記第2のセンサが前記輝度変動イベントに対して前記第2のイベント信号を生成したときに、前記第3のセンサが前記輝度変動イベントに対して第3のイベント信号を生成したか否かによって前記被写体の階調を判定する、請求項2に記載のセンサシステム。 - 第1の画素アドレスにおける光量の変動を第1の感度で検出する第1のセンサが輝度変動イベントに対して生成した第1のイベント信号に、前記第1の画素アドレスに隣接または重複する第2の画素アドレスにおける光量の変動を前記第1の感度よりも低い第2の感度で検出する第2のセンサが前記輝度変動イベントに対して生成した第2のイベント信号を関連付けるイベント信号関連付け部と、
前記第1のイベント信号に関連付けられる前記第2のイベント信号の有無に応じて前記輝度変動イベントを発生させた被写体の階調を判定する階調判定部と
を備える画像処理装置。 - 前記イベント信号関連付け部は、前記第1のイベント信号に、前記第1の画素アドレスに隣接または重複する第3の画素アドレスにおける光量の変動を前記第2の感度よりも低い第3の感度で検出する第3のセンサが前記輝度変動イベントに対して生成した第3のイベント信号を関連付け、
前記階調判定部は、前記第1のイベント信号に関連付けられる前記第3のイベント信号の有無に応じて前記被写体の階調を判定する、請求項4に記載の画像処理装置。 - 第1の画素アドレスにおける光量の変動を第1の感度で検出する第1のセンサが、輝度変動イベントに対して第1のイベント信号を生成するステップと、
前記第1の画素アドレスに隣接または重複する第2の画素アドレスにおける光量の変動を前記第1の感度よりも低い第2の感度で検出する第2のセンサが前記輝度変動イベントに対して第2のイベント信号を生成するか、または前記第2のイベント信号を生成しないステップと、
前記第1のイベント信号に前記第2のイベント信号を関連付けるステップと、
前記第1のイベント信号に関連付けられる前記第2のイベント信号の有無に応じて前記輝度変動イベントを発生させた被写体の階調を判定するステップと
を含む画像処理方法。 - 第1の画素アドレスにおける光量の変動を第1の感度で検出する第1のセンサが輝度変動イベントに対して生成した第1のイベント信号に、前記第1の画素アドレスに隣接または重複する第2の画素アドレスにおける光量の変動を前記第1の感度よりも低い第2の感度で検出する第2のセンサが前記輝度変動イベントに対して生成した第2のイベント信号を関連付ける機能と、
前記第1のイベント信号に関連付けられる前記第2のイベント信号の有無に応じて前記輝度変動イベントを発生させた被写体の階調を判定する機能と
をコンピュータに実現させるためのプログラム。
Priority Applications (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2021541868A JP7191238B2 (ja) | 2019-08-28 | 2019-08-28 | センサシステム、画像処理装置、画像処理方法およびプログラム |
| US17/635,301 US11653109B2 (en) | 2019-08-28 | 2019-08-28 | Sensor system, image processing apparatus, image processing method, and program |
| KR1020227004213A KR102746170B1 (ko) | 2019-08-28 | 2019-08-28 | 센서 시스템, 화상 처리 장치, 화상 처리 방법 및 프로그램 |
| PCT/JP2019/033708 WO2021038751A1 (ja) | 2019-08-28 | 2019-08-28 | センサシステム、画像処理装置、画像処理方法およびプログラム |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/JP2019/033708 WO2021038751A1 (ja) | 2019-08-28 | 2019-08-28 | センサシステム、画像処理装置、画像処理方法およびプログラム |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2021038751A1 true WO2021038751A1 (ja) | 2021-03-04 |
Family
ID=74684377
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/JP2019/033708 Ceased WO2021038751A1 (ja) | 2019-08-28 | 2019-08-28 | センサシステム、画像処理装置、画像処理方法およびプログラム |
Country Status (4)
| Country | Link |
|---|---|
| US (1) | US11653109B2 (ja) |
| JP (1) | JP7191238B2 (ja) |
| KR (1) | KR102746170B1 (ja) |
| WO (1) | WO2021038751A1 (ja) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2018186478A (ja) * | 2017-04-25 | 2018-11-22 | ソニーセミコンダクタソリューションズ株式会社 | 固体撮像素子、撮像装置、および、固体撮像素子の制御方法 |
| WO2019135411A1 (ja) * | 2018-01-05 | 2019-07-11 | 株式会社ニコン | 検出装置およびセンサ |
| JP2019134271A (ja) * | 2018-01-31 | 2019-08-08 | ソニーセミコンダクタソリューションズ株式会社 | 固体撮像素子、撮像装置、および、固体撮像素子の制御方法 |
Family Cites Families (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2012105225A (ja) * | 2010-11-12 | 2012-05-31 | Sony Corp | 画像処理装置、撮像装置、および画像処理方法、並びにプログラム |
| JP2012257193A (ja) | 2011-05-13 | 2012-12-27 | Sony Corp | 画像処理装置、撮像装置、および画像処理方法、並びにプログラム |
| TWI644568B (zh) * | 2013-07-23 | 2018-12-11 | 新力股份有限公司 | 攝像元件、攝像方法及攝像程式 |
| JP6373577B2 (ja) * | 2013-12-25 | 2018-08-15 | ザインエレクトロニクス株式会社 | 撮像制御装置 |
| KR102136055B1 (ko) * | 2014-01-08 | 2020-07-21 | 삼성전자 주식회사 | 오픈-루프 증폭기를 포함하는 비전 센서 칩, 이의 동작 방법, 및 이를 포함하는 데이터 처리 시스템 |
| JP6299299B2 (ja) * | 2014-03-14 | 2018-03-28 | オムロン株式会社 | 事象検出装置および事象検出方法 |
| US10277805B2 (en) * | 2014-05-30 | 2019-04-30 | Hitachi Kokusai Electric Inc. | Monitoring system and camera device |
| FR3034204A1 (ja) | 2015-03-23 | 2016-09-30 | Stmicroelectronics (Grenoble 2) Sas | |
| KR102523136B1 (ko) * | 2015-09-01 | 2023-04-19 | 삼성전자주식회사 | 이벤트 기반 센서 및 이벤트 기반 센서의 픽셀 |
| KR102421141B1 (ko) | 2015-10-30 | 2022-07-14 | 삼성전자주식회사 | 이벤트 신호 및 영상의 저장 방법 및 저장 장치, 저장 장치로 이벤트 신호를 전송하는 비전 센서의 동작 방법 |
| CN107801425B (zh) * | 2016-03-31 | 2021-05-14 | 索尼公司 | 固态摄像器件及其驱动方法和电子设备 |
| JP2018022935A (ja) * | 2016-08-01 | 2018-02-08 | ソニー株式会社 | 撮像装置、および、撮像装置の制御方法 |
| KR102794753B1 (ko) * | 2016-11-24 | 2025-04-14 | 삼성전자주식회사 | 디스플레이 장치 및 그 제어 방법 |
| CN108574793B (zh) | 2017-03-08 | 2022-05-10 | 三星电子株式会社 | 被配置为重新生成时间戳的图像处理设备及包括其在内的电子设备 |
| US20190281238A1 (en) | 2018-03-09 | 2019-09-12 | Caeleste Cvba | Double source follower hdr pixel |
| US10598936B1 (en) | 2018-04-23 | 2020-03-24 | Facebook Technologies, Llc | Multi-mode active pixel sensor |
| US11463636B2 (en) * | 2018-06-27 | 2022-10-04 | Facebook Technologies, Llc | Pixel sensor having multiple photodiodes |
| EP3920211B1 (en) * | 2019-01-31 | 2025-10-15 | Sony Semiconductor Solutions Corporation | Solid-state imaging device, and imaging device |
| JP2020136958A (ja) * | 2019-02-21 | 2020-08-31 | ソニーセミコンダクタソリューションズ株式会社 | イベント信号検出センサ及び制御方法 |
| JP7352637B2 (ja) * | 2019-08-28 | 2023-09-28 | 株式会社ソニー・インタラクティブエンタテインメント | センサモジュール、センサシステム、画像処理装置、画像処理方法およびプログラム |
| JP7489189B2 (ja) * | 2019-08-30 | 2024-05-23 | ソニーセミコンダクタソリューションズ株式会社 | 固体撮像素子、撮像装置、および、固体撮像素子の制御方法 |
-
2019
- 2019-08-28 KR KR1020227004213A patent/KR102746170B1/ko active Active
- 2019-08-28 JP JP2021541868A patent/JP7191238B2/ja active Active
- 2019-08-28 US US17/635,301 patent/US11653109B2/en active Active
- 2019-08-28 WO PCT/JP2019/033708 patent/WO2021038751A1/ja not_active Ceased
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2018186478A (ja) * | 2017-04-25 | 2018-11-22 | ソニーセミコンダクタソリューションズ株式会社 | 固体撮像素子、撮像装置、および、固体撮像素子の制御方法 |
| WO2019135411A1 (ja) * | 2018-01-05 | 2019-07-11 | 株式会社ニコン | 検出装置およびセンサ |
| JP2019134271A (ja) * | 2018-01-31 | 2019-08-08 | ソニーセミコンダクタソリューションズ株式会社 | 固体撮像素子、撮像装置、および、固体撮像素子の制御方法 |
Also Published As
| Publication number | Publication date |
|---|---|
| KR102746170B1 (ko) | 2024-12-26 |
| JPWO2021038751A1 (ja) | 2021-03-04 |
| KR20220074854A (ko) | 2022-06-03 |
| JP7191238B2 (ja) | 2022-12-16 |
| US11653109B2 (en) | 2023-05-16 |
| US20220311957A1 (en) | 2022-09-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2021038749A1 (ja) | センサシステム、画像処理装置、画像処理方法およびプログラム | |
| KR102707749B1 (ko) | 옵티컬 블랙 영역을 이용하여 이벤트 신호들을 보정하도록 구성되는 다이나믹 비전 센서 및 그 동작 방법 | |
| US11678074B2 (en) | Sensor control apparatus, sensor control method, and program | |
| US11330219B2 (en) | Dynamic vision sensor system | |
| US8643751B2 (en) | Method for detecting dead pixels and computer program product thereof | |
| JP2013530466A (ja) | ステレオカメラシステムの光学式自己診断 | |
| KR20210006106A (ko) | 다이내믹 비젼 센서의 이벤트 보정 방법 및 이를 수행하는 이미지 센서 장치 | |
| CN109640011A (zh) | 固态成像设备和成像系统 | |
| US20100110209A1 (en) | Fast motion measurement device for gaming | |
| CN109903324B (zh) | 一种深度图像获取方法及装置 | |
| WO2019026457A1 (ja) | 画像監視装置、画像監視方法、画像監視プログラムおよび記録媒体 | |
| CN117135482A (zh) | 影像感测装置及其影像感测方法 | |
| US20240107194A1 (en) | Delay Equalization in Event-Based Vision Sensors | |
| EP2944257B1 (en) | Radiation imaging apparatus, method of determining radiation irradiation, and storage medium | |
| US8576310B2 (en) | Image processing apparatus, camera module, and image processing method | |
| US10628951B2 (en) | Distance measurement system applicable to different reflecting surfaces and computer system | |
| JP7191238B2 (ja) | センサシステム、画像処理装置、画像処理方法およびプログラム | |
| JP2009017158A (ja) | カメラ検査装置 | |
| US11546538B2 (en) | Photoelectric conversion apparatus, photoelectric conversion system, moving body, and method for checking photoelectric conversion apparatus | |
| KR20240165386A (ko) | 이미지 센서, 데이터 구조 | |
| JPWO2009107636A1 (ja) | 点滅信号検出回路、点滅信号検出方法、物体検知装置、物体検知方法、及び撮像システム | |
| US20260075335A1 (en) | Individual spad pixel control for mitigating hot pixels and reducing power consumption | |
| US8054347B2 (en) | Image sensor systems having improved noise performance | |
| JP2002330354A (ja) | 欠損画素検出補正装置、欠損画素検出補正方法、欠損画素検出補正プログラム、および、映像信号処理装置 | |
| US12456176B2 (en) | Image signal processor and noise removal method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 19943111 Country of ref document: EP Kind code of ref document: A1 |
|
| ENP | Entry into the national phase |
Ref document number: 2021541868 Country of ref document: JP Kind code of ref document: A |
|
| ENP | Entry into the national phase |
Ref document number: 20227004213 Country of ref document: KR Kind code of ref document: A |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 19943111 Country of ref document: EP Kind code of ref document: A1 |