Bionic vision chip based on complementary perception theory, and system on chip

Abstract

Provided in the present application are a bionic vision chip based on a complementary perception theory, and a system on a chip. The chip comprises a pixel array, a first channel and a second channel, wherein the pixel array consists of a first pixel group and a second pixel group, and perceptual properties of the first pixel group and the second pixel group are complementary; the perceptual properties at least comprise a primitive set; the first channel controls, reads and transmits a pixel signal of the first pixel group; and the second channel controls, reads and transmits a pixel signal of the second pixel group. In the present application, different visual properties are assigned to primitives with complementary characteristics, and a complementary vision chip architecture based on the theory is designed, such that imaging with a high dynamic range, a high speed, a high resolution and a high image quality can be realized with an extremely low bandwidth and power consumption, thereby breaking through the bandwidth wall and power consumption wall that present vision sensors are facing, and realizing an efficient and robust visual representation.

Description

A bionic vision chip and system-on-chip based on complementary perception theory

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims priority to Chinese patent application No. 2023114206723, filed on October 30, 2023, entitled “A Bionic Vision Chip and System on Chip Based on Complementary Perception Theory,” which is incorporated herein by reference in its entirety. Technical Field

[0003] The present application relates to the field of visual sensing technology, and in particular to a bionic vision chip and system-on-chip based on complementary perception theory. Background Art

[0004] A visual sensor refers to an instrument that uses optical elements and imaging devices to obtain image information of the external environment. Visual sensors in the prior art generally include: Active Pixel Sensor (APS) and Dynamic Vision Sensor (DVS). Among them, active pixel sensors are usually image sensors based on the frame sampling principle, which are widely used in the camera units of mobile phones or cameras. This type of image sensor has the advantages of high color reproduction and image quality. However, the dynamic range of the image signal it obtains is small, and the shooting speed is slow. The characteristic of dynamic vision sensors is that they can perceive changes in dynamic scenes in the form of sparse event streams. Because of the fast shooting speed and the large dynamic range of the image signal obtained, this type of sensor has the problems of low resolution and excessive loss of effective information.

[0005] The existing APS and DVS only implement a single type of pixel. DAVIS, which combines APS and DVS, can record both single-frame images and event information, and has the advantages of high spatial resolution of traditional cameras and high temporal resolution of DVS cameras. However, in DAVIS, although two types of pixels are implemented, the relationship between the two pixels is independent. In order to achieve high-performance sensing, too much cost is consumed, and information complementarity cannot be achieved. This results in the current artificial vision chips and systems being unable to achieve high-speed response, sparse response, high dynamic range, high resolution, and high image accuracy at a very low cost (bandwidth and power consumption) like the human eye, making it difficult to achieve efficient and robust visual representation.

[0006] Summary of the Invention

[0007] The present application provides a bionic vision chip and system-on-chip based on the theory of complementary perception, which is used to address the defects of the visual sensor chips in the existing technology that cannot achieve pixel information complementarity, have power consumption walls and bandwidth walls, and find it difficult to achieve efficient and robust visual representation. The present application allocates different visual properties as primitives with complementary characteristics, and designs a complementary vision chip architecture based on this theory, which can achieve high dynamic range, high speed, high resolution, and high image quality imaging with extremely low bandwidth and power consumption, breaking the bandwidth wall and power consumption wall faced by current visual sensors and achieving efficient and robust visual representation.

[0008] The present application provides a bionic vision chip based on complementary perception theory, comprising: a pixel array consisting of a first pixel group and a second pixel group, wherein the perceptual properties of the first pixel group and the second pixel group are complementary; the perceptual properties include at least a primitive set; the primitive set is one or more of temporal resolution, spatial resolution, receptive field, pixel size, sensitivity, spectral response range, response mode, and accuracy; a first path adapted to the first pixel group, for controlling, reading out, and transmitting pixel signals of the first pixel group; a second path adapted to the second pixel group, for controlling, reading out, and transmitting pixel signals of the second pixel group.

[0009] According to a bionic vision chip based on complementary perception theory provided by the present application, the first pixel group is a color visible light pixel group, and the second pixel group is one of a grayscale visible light pixel group, an infrared pixel group, and an ultraviolet pixel group; or, the first pixel group is one of an infrared pixel group and an ultraviolet pixel group, and the second pixel group is a grayscale visible light pixel group.

[0010] According to a bionic vision chip based on complementary perception theory provided by the present application, the first pixel group and the second pixel group are arranged separately.

[0011] According to a bionic vision chip based on complementary perception theory provided by the present application, the sub-pixels of the first pixel group are arranged between the sub-pixels of the second pixel group.

[0012] According to a bionic vision chip based on complementary perception theory provided by the present application, the sub-pixels of the first pixel group are arranged at the edges of the sub-pixels of the second pixel group.

[0013] According to a bionic vision chip based on complementary perception theory provided by the present application, the sub-pixels of the first pixel group alternate with the sub-pixels of the second pixel group in a row direction.

[0014] According to a bionic vision chip based on complementary perception theory provided by the present application, the sub-pixels of the first pixel group alternate with the sub-pixels of the second pixel group in a column direction.

[0015] According to a bionic vision chip based on complementary perception theory provided by the present application, the perceptual properties of the sub-pixels of the first pixel group and the sub-pixels of the second pixel group are complementary.

[0016] According to a bionic vision chip based on complementary perception theory provided by the present application, the first path is specifically used to control, read out and transmit the spatiotemporal differential pixel signals of the first pixel group, and the second path is specifically used to control, read out and transmit the absolute value pixel signals of the second pixel group.

[0017] The present application also provides a system on a chip, comprising the above-mentioned bionic vision chip based on complementary perception theory, and also comprising a processor for processing data output from the bionic vision chip based on complementary perception theory.

[0018] The present application provides a bionic vision chip and system-on-chip based on the theory of complementary perception, which includes a pixel array, a first path, and a second path. The pixel array is composed of a first pixel group and a second pixel group, and the perceptual properties of the first pixel group and the second pixel group are complementary; the perceptual properties include at least a primitive set; the primitive set is one or more of temporal resolution, spatial resolution, receptive field, pixel size, sensitivity, spectral response range, response mode, and accuracy; the first path controls, reads out, and transmits pixel signals of the first pixel group; the second path controls, reads out, and transmits pixel signals of the second pixel group. The present application assigns different visual properties to primitives with complementary characteristics, and designs a complementary vision chip architecture based on this theory, which can achieve high dynamic range, high speed, high resolution, and high image quality imaging with extremely low bandwidth and power consumption, breaking the bandwidth wall and power consumption wall faced by current visual sensors and achieving efficient and robust visual representation. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the present application or the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0020] FIG1 is a schematic diagram of a pixel array in a bionic vision chip based on complementary perception theory provided in an embodiment of the present application;

[0021] FIG2 is a schematic diagram of a pixel array in another bionic vision chip based on complementary perception theory provided in an embodiment of the present application;

[0022] FIG3 is a schematic diagram of a pixel array in another bionic vision chip based on complementary perception theory provided in an embodiment of the present application;

[0023] FIG4 is a schematic diagram of a first pathway and a second pathway in a bionic vision chip based on complementary perception theory provided in an embodiment of the present application;

[0024] FIG5 is a schematic diagram of the first pathway and the second pathway in another bionic vision chip based on complementary perception theory provided in an embodiment of the present application. DETAILED DESCRIPTION

[0025] To make the objectives, technical solutions, and advantages of this application more clear, the technical solutions of this application will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of this application.

[0026] The basic principle of current mainstream image sensors is frame-based shooting and recording, achieved through an active pixel (APS) array. APS sensors can only process color images arranged in pixel matrix image frames. They offer advantages such as high color reproduction, high resolution, and high image quality. However, the dynamic range of the image signals they acquire is limited, and the capture speed is slow.

[0027] An event camera, also known as a dynamic visual sensor (DVS), is a new type of imaging system. Unlike traditional cameras, which use a shutter to control the frame rate and record light intensity on a per-frame basis for all pixels, an event camera is sensitive to the rate of change of light intensity. Each pixel independently records the change in the logarithm of the light intensity at that pixel, generating a positive or negative pulse when the change exceeds a threshold. The asynchronous nature of event cameras allows them to be unrestricted by shutter speeds and boasts extremely high temporal resolution (a frame rate of approximately 1,000,000 fps, compared to approximately 100 fps for traditional cameras). This, combined with their sensitivity to change, makes them naturally suited for tasks such as motion monitoring. Another type of camera, called DAVIS, combines a traditional active pixel sensor (APS) with a DVS, capable of recording both single-frame images and event information, combining the high spatial resolution of traditional cameras with the high temporal resolution of DVS cameras.

[0028] Existing DVS and APS systems only implement a single type of pixel. While DAVIS implements two types of pixels, the relationship between the two is independent. This consumes excessive effort to achieve high-performance sensing, and prevents information complementarity. This results in current artificial vision chips and systems being unable to simultaneously achieve high-speed response, sparse response, high dynamic range, high resolution, and high image accuracy at the same low cost (bandwidth and power consumption) as the human eye. This is known as the power consumption wall and bandwidth wall of vision sensors.

[0029] In order to solve the technical problems existing in the prior art, the present application provides a bionic vision chip based on complementary perception theory, including: a complementary perception pixel array composed of a first pixel group and a second pixel group, the perception properties of the first pixel group and the second pixel group are complementary; the perception properties include at least a primitive set; the primitive set is one or more of temporal resolution, spatial resolution, receptive field, pixel size, sensitivity, spectral response range, response mode, and accuracy; a first path adapted to the first pixel group, for controlling, reading out and transmitting pixel signals of the first pixel group; a second path adapted to the second pixel group, for controlling, reading out and transmitting pixel signals of the second pixel group.

[0030] The human visual system is currently the most comprehensive visual system, capable of achieving dynamic range, speed, resolution, and image quality far exceeding those of artificial vision systems at a fraction of the bandwidth and power consumption. The most important reason for this is that humans decompose visual information into complementary streams, processing them in parallel and asynchronously in the brain. This allows for efficient responses to both fine static information and rapid dynamic changes. For example, the human visual system can rapidly identify moving targets, whether at noon or dusk, in open scenes or partially obscured. This results in robustness and versatility far exceeding existing computer vision systems.

[0031] This application is inspired by human vision and proposes a complementary visual perception theory and data structure. The complementary theory refers to the characteristics of human vision and requires that the acquired data belong to multiple different paths, and the data structure of each path should be complementary in different properties. These properties are called primitives. Typical primitives include temporal resolution (frame rate, speed, delay fast and slow complementarity), spatial resolution (high and low complementarity), receptive field (large and small complementarity), pixel size (large and small complementarity), sensitivity (high and low complementarity), spectral response range (range complementarity of quantum efficiency), response mode (integrated intensity quantity, differential change complementarity), precision (high and low complementarity), etc. Primitives can also be extended to other visual perception properties. These primitives can be freely combined into different perceptual physical paths (i.e., the physical realization of perception). By rationally allocating complementary primitives to the paths, such as high speed and low resolution + high resolution and low speed, while balancing the bandwidth and power consumption of each path, extremely low total bandwidth and total power consumption are ultimately achieved.

[0032] Based on this theory, this application proposes a responsive sensing chip architecture, including a complementary sensing pixel array, control and readout path architecture.

[0033] Based on the aforementioned visual perception data structure, this application proposes a perceptual pixel array with complementary pixel arrangement format, size, resolution, sensitivity, and spectrum in various spatial and single-point dimensions to support the proposed data structure. This array can achieve both ultra-high resolution and dynamic range at a very low cost.

[0034] To achieve high-speed and sparse response, this application implements a control, information readout, and transmission path adapted to a complementary pixel array. This transmission backend achieves complementary information response mode, speed, and accuracy, further supporting high dynamic range while achieving ultra-high-speed response.

[0035] The present application can flexibly complement the different primitives of the sub-pixels of the first pixel group and the sub-pixels of the second pixel group, including spectral response (white spectrum corresponds to grayscale, red, green and blue correspond to the visible light spectrum, infrared and ultraviolet correspond to the non-visible spectrum), sensitivity, saturation value, noise, size, receptive field, speed, resolution, etc., thereby realizing the decomposition of different dimensions of optical visual information, achieving ultra-high resolution, ultra-high dynamic range, extremely low noise, wide spectral response, multiple receptive fields, etc.

[0036] The bionic vision chip based on the theory of complementary perception can be used to capture a target object to acquire an image signal or a video signal. The target object can be a static person, a dynamic person, a static scene, a dynamic scene, or other objects, and the present application is not limited thereto. The bionic vision chip can be applied to small consumer electronic products such as scanning pens.

[0037] The pixel may include a photosensitive device, such as a photodiode (PD), which can convert light signals into corresponding electrical signals. The photosensitive device can be a UV-sensitive device, such as a UV photodiode. In this way, the bionic vision chip can sense the intensity changes of ultraviolet light in the target light signal, such as the ultraviolet band of 760nm to 1mm.

[0038] Based on the above embodiment:

[0039] As a preferred embodiment, the first pixel group is a color visible light pixel group, and the second pixel group is one of a grayscale visible light pixel group, an infrared pixel group, and an ultraviolet pixel group; or, the first pixel group is one of an infrared pixel group and an ultraviolet pixel group, and the second pixel group is a grayscale visible light pixel group.

[0040] A complementary sensing pixel array refers to an array of pixels arranged on a bionic vision chip. The array can have various shapes, such as rectangular, diamond, or triangular. The complementary sensing pixel array receives light signals and converts them into electrical signals. The complementary sensing pixel array can be a combination of one or more of the following: color pixels, panchromatic pixels, infrared pixels, and ultraviolet pixels.

[0041] A color pixel is a pixel that receives colored light (e.g., red, green, and blue) and converts the optical signal into an electrical signal. These pixels include red, green, and blue pixels. A color pixel can include a color filter, such as a red filter, a green filter, or a blue filter. A color filter separates light, allowing only the corresponding colored light to pass through. This allows the color pixel to receive only the light of the color corresponding to the color filter. For example, a red pixel with a red filter will only receive red light signals. Red (R) can correspond to a color in the wavelength range of 600nm to 750nm (inclusive), green (G) can correspond to a color in the wavelength range of 495nm to 570nm (inclusive), and blue (B) can correspond to a color in the wavelength range of 450nm to 495nm (inclusive).

[0042] A full-color pixel refers to a pixel point that is used to receive white light and convert the optical signal into an electrical signal. A full-color pixel can also be called a white pixel. A full-color pixel includes a full-color filter, which can also be called a white filter, a transparent layer, or a transparent filter. A full-color filter allows visible light to pass through, or in other words, a full-color filter does not separate the color of light, so that a full-color pixel can receive any visible light. Among them, the range of visible light can be light with a wavelength of 400nm to 800nm. At the same time, compared with color pixels, full-color pixels do not filter light, so full-color pixels usually have a higher amount of light input than color pixels.

[0043] The color correspondence between the first pixel group and the second pixel group can be in the following forms:

[0044] The first pixel group may be a color visible light pixel group, and the second pixel group may be a grayscale visible light pixel group;

[0045] The first pixel group may be a color visible light pixel group, and the second pixel group may be an infrared pixel group;

[0046] The first pixel group may be a color visible light pixel group, and the second pixel group may be an ultraviolet pixel group;

[0047] The first pixel group may be an infrared pixel group, and the second pixel group may be a grayscale visible light pixel group;

[0048] The first pixel group may be an ultraviolet pixel group, and the second pixel group may be a grayscale visible light pixel group.

[0049] Please refer to Figure 1, which is a schematic diagram of a pixel array in a bionic vision chip based on complementary perception theory provided in an embodiment of the present application.

[0050] Grayscale pixels are pixels that respond to white spectrum grayscale, have high saturation value, large size, and high speed; color pixels are pixels that respond to color, have high sensitivity, and high resolution.

[0051] Please refer to Figure 2, which is a schematic diagram of a pixel array in another bionic vision chip based on complementary perception theory provided in an embodiment of the present application.

[0052] Color pixels are pixels with color response, high saturation value, large size and high speed; grayscale pixels are pixels with white spectrum grayscale response, high sensitivity and high resolution.

[0053] Please refer to Figure 3, which is a schematic diagram of a pixel array in another bionic vision chip based on complementary perception theory provided in an embodiment of the present application.

[0054] Color pixels are pixels that respond to visible light colors, have high saturation values, large sizes, and high speeds; grayscale pixels are pixels that respond to white spectrum grayscales, have high sensitivity, and high resolutions; and ultraviolet and infrared pixels are pixels that respond to non-visible light colors.

[0055] As a preferred embodiment, the first pixel group and the second pixel group are arranged separately.

[0056] As a preferred embodiment, the sub-pixels of the first pixel group are arranged between the sub-pixels of the second pixel group.

[0057] A plurality of sub-pixels of the first pixel group may be placed between adjacent sub-pixels of the second pixel group.

[0058] The sub-pixels of the first pixel group and the sub-pixels of the second pixel group may be arranged at unequal intervals.

[0059] As a preferred embodiment, the sub-pixels of the first pixel group are arranged at edges of the sub-pixels of the second pixel group.

[0060] The pixel array may include a second pixel group positioned inside and a first pixel group positioned at a boundary of the second pixel group.

[0061] The sub-pixels of the first pixel group may be arranged at edges of the sub-pixels of the second pixel group. The sub-pixels of the second pixel group may be arranged at edges of the sub-pixels of the first pixel group.

[0062] The sub-pixels of the first pixel group may be arranged in a single line at the boundary of the sub-pixels of the second pixel group, or the sub-pixels of the first pixel group may be arranged in a multi-line form.

[0063] As a preferred embodiment, the sub-pixels of the first pixel group alternate with the sub-pixels of the second pixel group in the row direction.

[0064] The pixel array may include only sub-pixels of the first pixel group or only sub-pixels of the second pixel group in each column.

[0065] Although a column of sub-pixels of the first pixel group alternates with a column of sub-pixels of the second pixel group, multiple columns of sub-pixels of the first pixel group may be placed between adjacent columns of sub-pixels of the second pixel group.

[0066] The sub-pixels of the first pixel group in each column and the sub-pixels of the second pixel group in each column may be arranged at unequal intervals.

[0067] As a preferred embodiment, the sub-pixels of the first pixel group alternate with the sub-pixels of the second pixel group in a column direction.

[0068] The pixel array may include only sub-pixels of the first pixel group or only sub-pixels of the second pixel group in each row.

[0069] Although a row of subpixels of the first pixel group alternates with a row of subpixels of the second pixel group, multiple rows of subpixels of the first pixel group may be placed between adjacent rows of subpixels of the second pixel group.

[0070] The sub-pixels of the first pixel group in each row and the sub-pixels of the second pixel group in each row may be arranged at unequal intervals.

[0071] As a preferred embodiment, the sub-pixels of the first pixel group and the sub-pixels of the second pixel group have different sizes.

[0072] The sub-pixels of the first pixel group and the sub-pixels of the second pixel group have the same size. Since the sub-pixels of the first pixel group and the sub-pixels of the second pixel group have the same size, when the pixel array has a 2×2 array form, the pixel array can include one sub-pixel of the first pixel group and three sub-pixels of the second pixel group.

[0073] One subpixel of the first pixel group has the same size as four subpixels of the second pixel group. Since one subpixel of the first pixel group has the same size as four subpixels of the second pixel group, when the pixel array has a 3×3 array form, the pixel array can include one subpixel of the first pixel group and five subpixels of the second pixel group.

[0074] The relative sizes of the sub-pixels of the first pixel group and the sub-pixels of the second pixel group may be varied in various ways.

[0075] As a preferred embodiment, the first path is specifically used to control, read out and transmit the spatiotemporal differential pixel signal of the first pixel group, and the second path is specifically used to control, read out and transmit the absolute value pixel signal of the second pixel group.

[0076] With the help of high-performance signal readout channels, this application can achieve the complementarity of primitives such as response modality, accuracy, and speed, and further support the acquisition of visual information with high dynamic range, high speed, high resolution, and high image quality.

[0077] The present application can flexibly allocate the two types of complementary pixels to two different pathways. The first pathway is used to respond to sparse, high-speed spatiotemporal variations (i.e., spatiotemporal differential pixel signals), while the second pathway is mainly responsible for high-precision absolute intensity quantities (i.e., absolute value pixel signals).

[0078] In the first path, the differential signal reading circuit reads the grayscale pixel value by responding only to the change in value, while the second path reads only its absolute value. This complements the response modalities, enabling the perception of both static, detailed images and dynamic, spatiotemporal information in video streams.

[0079] Different paths have different speeds and accuracies. The first path is responsible for high-speed and sparse spatiotemporal variations, so it uses slightly lower-precision sampling and transmission to ensure low bandwidth and low power consumption. Meanwhile, the second path uses a high-precision sampling mode to achieve higher image quality and resolution.

[0080] Here, two paths are used as an example to illustrate the typical approach to path allocation. In reality, these primitives can be freely combined onto different perceptual physical paths (i.e., the physical implementation of perception). The proper allocation of primitives is controlled by calculating the total bandwidth and power consumption of all paths.

[0081] Please refer to Figure 4, which is a schematic diagram of the first pathway and the second pathway in a bionic vision chip based on complementary perception theory provided in an embodiment of the present application.

[0082] The first channel corresponds to TD+SD pixels, responding to high-speed sparse spatiotemporal changes, and the second channel corresponds to I pixels, responding to high-precision intensity absolute values.

[0083] Please refer to Figure 5, which is a schematic diagram of the first pathway and the second pathway in another bionic vision chip based on complementary perception theory provided in an embodiment of the present application.

[0084] The first channel corresponds to D pixels, responding to high-speed sparse spatiotemporal changes, and the second channel corresponds to intensity absolute value pixels, responding to high-precision intensity absolute values.

[0085] In summary, this application proposes a theory and data structure inspired by human vision, primarily by assigning different visual properties to primitives with complementary characteristics. Based on this theory, a complementary vision chip architecture is designed, which achieves complementarity among all visual primitives, including spectral response, sensitivity, saturation value, noise, size, receptive field, speed, resolution, accuracy, and response modality. This enables high-performance imaging with high dynamic range, high speed, high resolution, and high image quality at extremely low bandwidth and power consumption, breaking the bandwidth and power consumption barriers currently faced by visual sensors and achieving efficient and robust visual representation.

[0086] The present application also provides a system on a chip, comprising the above-mentioned bionic vision chip based on complementary perception theory, and also comprising a processor for processing data output from the bionic vision chip based on complementary perception theory.

[0087] For an introduction to a system on chip provided by this application, please refer to the above chip embodiment, and this application will not go into details here.

[0088] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A bionic visual chip based on complementary perception theory, comprising: A complementary perceptual pixel array consisting of a first pixel group and a second pixel group, wherein the perceptual properties of the first pixel group and the second pixel group are complementary; the perceptual properties include at least a primitive set; The primitive set is one or more of temporal resolution, spatial resolution, receptive field, pixel size, sensitivity, spectral response range, response mode, and accuracy; a first channel adapted to the first pixel group, for controlling, reading out and transmitting pixel signals of the first pixel group; A second path adapted to the second pixel group is used for controlling, reading out and transmitting pixel signals of the second pixel group.

2. The bionic visual chip based on complementary perception theory according to claim 1, wherein: The first pixel group is a color visible light pixel group, and the second pixel group is one of a grayscale visible light pixel group, an infrared pixel group, and an ultraviolet pixel group; or, The first pixel group is one of an infrared pixel group and an ultraviolet pixel group, and the second pixel group is a grayscale visible light pixel group.

3. The bionic visual chip based on complementary perception theory according to claim 1, wherein: The first pixel group is arranged separately from the second pixel group.

4. The bionic visual chip based on complementary perception theory according to claim 1, wherein: The sub-pixels of the first pixel group are arranged between the sub-pixels of the second pixel group.

5. The bionic visual chip based on complementary perception theory according to claim 1, wherein: The sub-pixels of the first pixel group are arranged at edges of the sub-pixels of the second pixel group.

6. The bionic visual chip based on complementary perception theory according to claim 1, wherein: The sub-pixels of the first pixel group alternate with the sub-pixels of the second pixel group in a row direction.

7. The bionic visual chip based on complementary perception theory according to claim 1, wherein: The sub-pixels of the first pixel group alternate with the sub-pixels of the second pixel group in a column direction.

8. The bionic visual chip based on complementary perception theory according to claim 1, wherein: The sub-pixels of the first pixel group have complementary perceptual properties to the sub-pixels of the second pixel group.

9. The bionic visual chip based on complementary perception theory according to any one of claims 1 to 8, wherein: The first path is specifically used to control, read out and transmit the spatiotemporal differential pixel signal of the first pixel group, and the second path is specifically used to control, read out and transmit the absolute value pixel signal of the second pixel group.

10. A system on chip, comprising the bionic vision chip based on complementary perception theory as claimed in any one of claims 1 to 9, and further comprising a processor for processing data output from the bionic vision chip based on complementary perception theory.

Images