Object Recognition Device Based on Optical Computing

Abstract

The present application discloses an object recognition device, which includes a first 4f optical system, a reference optical system, a first photosensitive detector, a second photosensitive detector, and a comparator; the first 4f optical system is configured to output a first optical signal with a first light intensity according to a first input object, wherein the first input object includes a target object; the first photosensitive detector is configured to convert the first optical signal into a first electrical signal; the reference optical system is configured to output a second optical signal with a second light intensity according to the first input object, and the second light intensity is not equal to the first light intensity; the second photosensitive detector is configured to convert the second optical signal into a second electrical signal; the comparator is configured to compare the first electrical signal and the second electrical signal, and output an identification result indicating that the target object has been identified according to the comparison result of the first electrical signal and the second electrical signal.

Description

Technical Field

[0001] The present invention relates to the field of optics, and in particular to an object recognition device based on optical computing. Background Art

[0002] Object recognition has a wide range of applications in various electronic devices. For example, in a mobile terminal (such as a mobile phone), various operations can be performed by recognizing whether a target object (such as a human face) exists. For instance, in a reading interface, when a human face is recognized, the screen can be kept constantly on, facilitating the user's reading.

[0003] When performing object recognition in the prior art, basically, an image of the target is collected through a camera, and then the recognition process is carried out by a processor built in the electronic device. However, since the camera consumes power during operation, and at the same time, the processor needs to perform a large number of operations during the recognition process, which also consumes a lot of power. Therefore, the overall power consumption of the electronic device is relatively large, especially for some scenarios where object recognition needs to be turned on for a long time, the power consumption will be even greater, thus affecting the battery life of the electronic device. Summary of the Invention

[0004] An embodiment of the present invention provides an object recognition device for solving the problem of excessive power consumption when performing object recognition in the prior art.

[0005] In a first aspect, the present application discloses an object recognition device, including a first 4f optical system, a reference optical system, a first photosensitive detector, a second photosensitive detector, and a comparator; the first 4f optical system is configured to output a first optical signal with a first light intensity according to a first input object, where the first input object includes a target object; the first photosensitive detector is configured to convert the first optical signal into a first electrical signal; the reference optical system is configured to output a second optical signal with a second light intensity according to the first input object, and the second light intensity is not equal to the first light intensity; the second photosensitive detector is configured to convert the second optical signal into a second electrical signal; the comparator is configured to compare the first electrical signal and the second electrical signal, and output an identification result indicating that the target object has been recognized according to the comparison result of the first electrical signal and the second electrical signal. In this embodiment, introducing the reference optical system can eliminate the interference caused by environmental factors, provide a reference for the object recognition of the first 4f optical system, so that object recognition can be performed more accurately. In addition, the optical system does not consume electricity, and the photosensitive detector and the comparator can be easily implemented with circuits with very low power consumption. Therefore, compared with the prior art, the power is reduced.

[0006] In a possible implementation of the first aspect, the relative difference between the first light intensity and the second light intensity is greater than or equal to a first preset threshold. For example, it can be greater than 50%. The relative difference can be expressed by the following formula: (A - B) / B, or (A - B) / (A + B), or (A - B) / A, etc., where A represents the magnitude of the first electrical signal and B represents the magnitude of the second electrical signal. When the relative difference reaches a certain threshold and is compared through the relative difference, it is less affected by the environment, can be identified more accurately, and is easier to implement.

[0007] In a possible implementation of the first aspect, the first 4f optical system is further configured to output a third optical signal with a third light intensity according to a second input object, where the second input object does not include the target object; the first photosensitive detector is further configured to convert the third optical signal into a third electrical signal; the comparison optical system is further configured to output a fourth optical signal with a fourth light intensity according to the second input object; the fourth light intensity is not equal to the third light intensity; the second photosensitive detector is further configured to convert the fourth optical signal into a fourth electrical signal; the comparator is further configured to compare the third electrical signal and the fourth electrical signal, and output an identification result of not identifying the target object according to the comparison result of the third electrical signal and the fourth electrical signal. Through this implementation, since the identification result of not identifying the object is output when the input object does not include the target object, misidentification can be avoided.

[0008] In a possible implementation of the first aspect, the relative difference between the third light intensity and the fourth light intensity is less than a second preset threshold. This threshold can be the same as the aforementioned first preset threshold, or it can be smaller. By comparing through the relative difference, it is less affected by the environment, can be identified more accurately, and is easier to implement.

[0009] In a possible implementation of the first aspect, the first 4f optical system is a 4f optical system with a first convolution operation ability. Through this 4f optical system, convolution operation can be realized, thereby performing object recognition.

[0010] In a possible implementation of the first aspect, the first 4f optical system includes: a first optical lens group, a first phase plate, and a second optical lens group arranged in sequence along the optical path; the first optical lens group includes one or more first optical lenses, and the focal length of the first optical lens group is f; the second optical lens group includes one or more second optical lenses, and the focal length of the second optical lens group is f; the distance between the first phase plate and the first optical lens group and the distance between the first phase plate and the second optical lens group are both f; the first phase plate is used to perform the first convolution operation. The above structure of the first 4f optical system is based on the structure of the existing 4f optical system and uses a phase plate to realize convolution operation, which is simple to implement.

[0011] In a possible implementation of the first aspect, the first phase plate has at least one of the following two characteristics: each part of the first phase plate has a thickness that conforms to the first convolution operation, and each part of the first phase plate has a light transmittance that conforms to the first convolution operation. When only any one of the above two characteristics is present, the convolution operation can be achieved. When both are present, it is equivalent to providing more parameters for the first convolution operation. Therefore, the performance of the convolution kernel is better, and the final recognition effect is also better.

[0012] In a possible implementation of the first aspect, the first 4f optical system includes: a first optical mirror group, a first phase plate, and a second optical mirror group arranged in sequence along the optical path; the first optical mirror group includes one or more first optical lenses, and the focal length of the first optical mirror group is f; the second optical mirror group includes one or more second optical lenses, and the focal length of the second optical mirror group is f; the distance between the first phase plate and the first optical mirror group and the distance between the first phase plate and the second optical mirror group are both f; the first phase plate has a plurality of partitions, and each partition is used to perform a convolution operation on a specific target object. Among them, each partition has at least one of the following two characteristics: each part in the partition has a thickness that conforms to the convolution operation corresponding to the partition, and each part in the partition has a light transmittance that conforms to the convolution operation corresponding to the partition. By setting different partitions, each partition can be trained separately for a similar sample to obtain the final phase plate. This training method has a higher final recognition rate compared to the method of training multiple target objects together; the entire first 4f optical system can output a first optical signal when the target object is recognized in any partition, thereby improving the recognition rate.

[0013] In a possible implementation of the first aspect, the reference optical system includes a second 4f optical system with the ability of a second convolution operation. The reference optical system is also a 4f optical system with the ability of convolution operation, which can make the recognition performance of the entire device better. Because even if one of the 4f optical systems is not very accurate in recognition, as long as the other 4f optical system can accurately recognize, then the difference between the two will meet a certain threshold, so as to obtain a correct recognition result.

[0014] In a possible implementation of the first aspect, the second 4f optical system includes: a third optical lens group, a second phase plate, and a fourth optical lens group arranged in sequence along the optical path; the third optical lens group includes one or more third optical lenses, and the focal length of the third optical lens group is f; the fourth optical lens group includes one or more fourth optical lenses, and the focal length of the fourth optical lens group is f; the distances between the second phase plate and the third optical lens group and between the second phase plate and the fourth optical lens group are both f; the second phase plate is used to perform the second convolution operation. The above structure of the second 4f optical system is based on the structure of the existing 4f optical system and uses a phase plate to implement the convolution operation, which is simple to implement.

[0015] In a possible implementation of the first aspect, the second phase plate has at least one of the following two characteristics: each part of the second phase plate has a thickness conforming to the second convolution operation, and each part of the second phase plate has a light transmittance conforming to the second convolution operation. When any one of the above two characteristics is present, the convolution operation can be realized. When both are present, it is equivalent to providing more parameters for the second convolution operation. Therefore, the performance of the convolution kernel is better and the final recognition effect is also better.

[0016] In a possible implementation of the first aspect, the second 4f optical system includes: a third optical lens group, a second phase plate, and a fourth optical lens group arranged in sequence along the optical path; the third optical lens group includes one or more third optical lenses, and the focal length of the third optical lens group is f; the fourth optical lens group includes one or more fourth optical lenses, and the focal length of the fourth optical lens group is f; the distances between the second phase plate and the third optical lens group and between the second phase plate and the fourth optical lens group are both f; the second phase plate has a plurality of partitions, and each partition is used to perform a convolution operation on a specific target object. Among them, each partition has at least one of the following two characteristics: each part in the partition has a thickness conforming to the convolution operation corresponding to the partition, and each part in the partition has a light transmittance conforming to the convolution operation corresponding to the partition. By setting different partitions, each partition can be trained separately for a similar sample to obtain the final phase plate. This training method has a higher recognition rate compared to the method of training multiple target objects together; the entire second 4f optical system can output a second optical signal when the target object is recognized in any partition, thereby improving the recognition rate.

[0017] In a possible implementation of the first aspect, the reference optical system includes a light intensity averaging device; the light intensity averaging device is used to average the intensity of the ambient light and output it to the second photosensitive detector. Using a light intensity averaging device is simple to implement and reduces the implementation and production costs.

[0018] In another possible implementation of the first aspect, the reference optical system does not use any optical devices, so that the input ambient light directly passes through the reference optical system and is then detected by the second photosensitive detector. In this way, the implementation is simpler.

[0019] In a possible implementation of the first aspect, the device further includes a beam splitter for splitting the input optical signal into two paths and respectively outputting them to the first 4f optical system and the reference optical system. Using a beam splitter can reduce the light input port of the device, thus better meeting the design requirements of the product.

[0020] In a second aspect, the present application discloses an electronic device, including a processing circuit and the object recognition device as described in the first aspect and various implementations of the first aspect. Wherein, the processing circuit is configured to process based on the recognition result output by the object recognition device.

[0021] In a possible implementation of the second aspect, the electronic device further includes a display screen. When the processing circuit is configured to process based on the recognition result output by the object recognition device, it is specifically configured to determine whether to turn off the display of the display screen based on the recognition result.

[0022] In a third aspect, the present application discloses an optical component, which includes the first 4f optical system, the reference optical system, the first photosensitive detector, and the second photosensitive detector of the object recognition device as described in the first aspect and various implementations of the first aspect. Specifically, the optical component includes a first 4f optical system, a reference optical system, a first photosensitive detector, and a second photosensitive detector; the first 4f optical system is configured to output a first optical signal with a first light intensity according to a first input object, and the first input object includes a target object; the first photosensitive detector is configured to convert the first optical signal into a first electrical signal; the reference optical system is configured to output a second optical signal with a second light intensity according to the first input object; the second light intensity is not equal to the first light intensity; the second photosensitive detector is configured to convert the second optical signal into a second electrical signal; wherein, when the first electrical signal and the second electrical signal are compared, the comparison result obtained through the comparison is used to determine that the first input object contains the target object.

[0023] In another possible implementation of the third aspect, the first 4f optical system is further configured to output a third optical signal with a third light intensity according to a second input object, where the second input object does not include the target object; the first photosensitive detector is further configured to convert the third optical signal into a third electrical signal; the reference optical system is further configured to output a fourth optical signal with a fourth light intensity according to the input object; the fourth light intensity is not equal to the third light intensity; the second photosensitive detector is further configured to convert the fourth optical signal into a fourth electrical signal; wherein, when the third electrical signal and the fourth electrical signal are compared, the comparison result obtained through the comparison is used to determine that the second input object does not contain the target object.

[0024] For example, the comparison can be performed based on the aforementioned comparator, or the comparison of signals can also be performed based on other circuits (such as a processor).

[0025] In another possible implementation of the third aspect, it further includes a beam splitter as described in an implementation of the first aspect.

[0026] For the specific implementation and effects of the third aspect and various possible implementations in the third aspect, reference can be made to the first aspect and various implementations in the first aspect, which will not be elaborated here.

[0027] In a fourth aspect, the present application discloses an optical component, which includes any two or any three combinations of the first 4f optical system, the reference optical system, the first photosensitive detector, and the second photosensitive detector of the object recognition device as described in the first aspect and various implementations of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0029] Figure 1 It is a schematic structural diagram of a basic 4f optical system in the prior art;

[0030] Figure 2 It is a schematic structural diagram of the object recognition device in Embodiment 1 of the present application;

[0031] Figure 3 It is a schematic structural diagram of the object recognition device in Embodiment 3 of the present application, where the reference optical system includes a 4f optical system;

[0032] Figure 4Schematic diagram of the object recognition device in Embodiment 3 of the present application, where the control optical system includes a light intensity averaging device;

[0033] Figure 5 Schematic diagram of the object recognition device in Embodiment 3 of the present application, where the control optical system does not use an optical device;

[0034] Figure 6 Schematic diagram of a phase plate in Embodiment 4 of the present application;

[0035] Figure 7 Schematic diagram of the object recognition device in Embodiment 5 of the present application, where the optical path is non - linear;

[0036] Figure 8 Schematic diagram of the object recognition device in Embodiment 6 of the present application, which includes only one light inlet;

[0037] Figure 9 Schematic diagram of an electronic device in Embodiment 8 of the present application. Detailed implementation manners

[0038] The following describes each embodiment of the present invention in conjunction with the respective drawings.

[0039] Embodiment 1

[0040] Refer to Figure 2 , this embodiment discloses an object recognition device, which includes a first 4f optical system 21, a control optical system 22, a first photosensitive detector 23, a second photosensitive detector 24, and a comparator 25.

[0041] The first 4f optical system in the present application is a 4f optical system with the first convolution operation ability and can perform object recognition. This 4f optical system with convolution operation function can be obtained based on the basic 4f optical system. The following specifically introduces the first 4f optical system. Refer to Figure 1 , which is a basic 4f optical system and can be used in optical branches such as Fourier optics. The basic 4f optical system includes two lenses, which can be convex lenses, concave lenses, or other lenses. The focal length of the lenses is f. In Figure 1In the figure, it is represented by lens 1 and lens 2. In the optical path of the 4f optical system, light rays sequentially pass through the object plane, lens 1, the spectral plane, and lens 2 and then are output to the image plane. That is, the optical signal on the image plane can be considered as the output of the 4f optical system. Among them, the distance between the object plane and lens 1 is f, the distance between the spectral plane and lens 1 is f, and the distances between lens 2 and the spectral plane and the image plane are also f. Therefore, it is called a "4f optical system". It should be noted that in the 4f optical system, the lens can also be replaced by a lens group (including multiple lenses to achieve the same function). At the same time, for different applications, different devices can be set on the spectral plane to achieve different functions.

[0042] In this application, by placing a modulation device with a modulation function on the spectral plane, the light can be modulated, thereby realizing the convolution operation function in the neural network and finally outputting the imaging result on the image plane to achieve object recognition. Among them, the device can be a device based on phase modulation, or a device based on light intensity modulation, or a device based on both phase and light intensity modulation. Specifically, the modulation device can be a phase plate. By training to manufacture a special phase plate, the first optical system can have the first convolution operation ability. By performing the first convolution operation on the input optical signal, the first optical signal is output for the target object (such as a human face). The specific implementation method of the phase plate is specifically introduced below (such as in Embodiment 4). For example, each part of the phase plate can have a thickness that conforms to the first convolution operation to achieve phase modulation, and / or each part of the phase plate can have a light transmittance that conforms to the first convolution operation to achieve light intensity modulation. In addition, it can also be achieved through a spatial light modulator (SLM). In this application, the first photosensitive detector can be located on the image plane to identify the light intensity of the image imaged on the image plane.

[0043] Specifically, the first 4f optical system is configured to output a first optical signal with a first optical intensity according to a first input object, where the first input object includes a target object; here, the input object refers to the object input into the system, that is, all objects that enter the first 4f optical system through optical signals. The prefix "first" is only used to distinguish from the names of different input objects that appear later and does not have other meanings. The first input object includes the target object and various background objects. The target object refers to the object to be recognized. For example, it can be an object (such as a human face, etc.) that meets certain rule requirements (such as a specific size). The background object refers to all objects other than the target object. For example, there may be various light sources, walls, desks, chairs, and other background objects around a person in a room. The light generated by these objects can be called background light (or ambient light). The first 4f optical system can output the first optical signal in a first target area, which is also the area where the optical signal output by the first 4f optical system is located. For example, it can be a part of the area located on the image plane. This part of the area can be used to place the first photosensitive device, so that the optical intensity in the first target area can be detected as the first optical intensity.

[0044] It should be noted that in this application, the first optical intensity is a general term for the optical signal output by the first 4f optical system and is not a fixed value. It can be understood that in different environments, the first optical intensity of the first optical signal will change. Similarly, other signals such as the second optical signal, the third optical signal, and the fourth optical signal in the following text do not represent a fixed optical signal but an optical signal that changes with the environment.

[0045] The first photosensitive detector is used to convert the first optical signal into a first electrical signal; this can facilitate subsequent processing by circuits based on electrical signals (such as comparators, processors, etc.).

[0046] The reference optical system is configured to output a second optical signal with a second optical intensity according to the first input object, and the second optical intensity is not equal to the first optical intensity.

[0047] In this application, the reference optical system can also be a 4f optical system with convolution operation ability like the first 4f optical system, or other optical systems.

[0048] The second photosensitive detector is used to convert the second optical intensity into a second electrical signal and output an identification result indicating that the target object has been identified according to the comparison result between the first electrical signal and the second electrical signal; where the second optical intensity is not equal to the first optical intensity.

[0049] The comparator is used to compare the first electrical signal and the second electrical signal, and output an identification result indicating that the target object has been identified according to the comparison result of the first electrical signal and the second electrical signal. Optionally, the relative difference or absolute difference of the magnitudes of the two electrical signals (such as the magnitudes of voltage or current) can be compared to output the identification result. Among them, the calculation of the relative difference refers to dividing the result of subtracting the two signals by one signal or by the sum of the two signals. For example, (A - B) / B, or (A - B) / (A + B), or (A - B) / A, etc., where A represents the magnitude of the first electrical signal and B represents the magnitude of the second electrical signal; when the relative difference reaches a certain threshold (such as greater than or equal to 30% or 50%), an object identification result (such as a specific level) is output to indicate that the target object has been identified. The comparator can perform a direct comparison, or can perform a comparison after some preprocessing (such as amplifying the signal). At this time, the signals for comparison are the signals after preprocessing.

[0050] The first 4f optical system is further configured to output a third optical signal with a third light intensity according to a second input object, where the second input object does not include the target object;

[0051] The first photosensitive detector is further configured to convert the third optical signal into a third electrical signal;

[0052] The reference optical system is further configured to output a fourth optical signal with a fourth light intensity according to the second input object; the fourth light intensity is not equal to the third light intensity;

[0053] The second photosensitive detector is further configured to convert the fourth optical signal into a fourth electrical signal;

[0054] The comparator is also used to compare the third electrical signal and the fourth electrical signal, and output an identification result of not identifying the target object according to the comparison result of the third electrical signal and the fourth electrical signal. For example, after the third optical signal and the fourth optical signal pass through the first photosensitive detector and the second photosensitive detector respectively, the third electrical signal and the fourth electrical signal are obtained. The relative difference between the magnitude of the third electrical signal and the magnitude of the fourth electrical signal is less than 50%. The comparator outputs an identification result of "not identifying the target object" (such as a specific level). It should be noted that the first preset threshold used when comparing the first electrical signal and the second electrical signal may be the same as or different from the first preset threshold used when comparing the third electrical signal and the fourth electrical signal. For example, 50% can be used as the threshold for both. When the relative difference between the magnitude of the first electrical signal and the magnitude of the second electrical signal is greater than or equal to 50%, it indicates that the input object includes the target object. When the relative difference between the magnitude of the third electrical signal and the magnitude of the fourth electrical signal is less than 50%, it indicates that the input object does not include the target. Or, different thresholds can also be used. For example, when the relative difference between the magnitude of the first electrical signal and the magnitude of the second electrical signal is greater than or equal to 50%, it indicates that the input object includes the target object. When the relative difference between the magnitude of the third electrical signal and the magnitude of the fourth electrical signal is less than or less than or equal to 40%, it indicates that the input object does not include the target. At the same time, it can be understood that since the two different scenarios (whether the input object contains the target object) are both judged by comparing with the threshold, in order to make the comparator perform two outputs for the two scenarios respectively, in one scenario, the difference between the two electrical signals can be greater than or equal to the first preset threshold, and in the other scenario, the difference between the two electrical signals can be less than or equal to (or less than) the second preset threshold.

[0055] In this embodiment, the first 4f optical system adopts an optical system with a convolution operation function, which can output a first optical signal with a first light intensity when the input object includes the target object. At the same time, the reference optical system can output a second optical signal with a second light intensity when the input object includes the target object. By comparing the intensities of the two optical signals (specifically, by converting them into electrical signals for comparison), the result of object recognition can be obtained. In this application, introducing the reference optical system can eliminate the interference caused by environmental factors and provide a benchmark for object recognition. It can be understood that the first optical signal output by the first 4f optical system is related to a certain environment. If the environment is changed, the light intensity of the optical signal will change. It is impossible to determine whether the difference in light intensity is caused by the environment or the failure to recognize the target object only based on the light intensity value of the first optical signal. By introducing the reference optical system, it is equivalent to providing a reference object for removing the environmental impact. Since the output of the reference system is also affected by the environment, when the two are compared, the part that is affected by the environment at the same time can be eliminated. In this way, the comparison result can be considered as the result of object recognition (such as the input object includes the target object). Similarly, the recognition result of not recognizing the target object can be output by comparing the third electrical signal and the fourth electrical signal.

[0056] At the same time, the optical system used in this embodiment does not consume electricity. The photosensitive detector and the comparator are very easy to implement with circuits with very low power consumption. Therefore, unlike the traditional method that requires processing by a processor with high power consumption, the processing power consumption is greatly reduced, which is particularly suitable for some scenarios that require continuous object recognition for a long time.

[0057] Embodiment 2

[0058] Based on the above embodiment, this embodiment specifically introduces the first 4f optical system. Refer to Figure 2 , in this embodiment, the first 4f optical system 21 includes: a first optical lens group 211, a first phase plate 212, and a second optical lens group 213 arranged in sequence along the optical path; where

[0059] The first optical lens group 211 includes one or more first optical lenses, and the focal length of the first optical lens group is f;

[0060] The second optical lens group 213 includes one or more second optical lenses, and the focal length of the second optical lens group is f;

[0061] In this application, in practice, the first optical lens group and the second optical lens group can each be composed of one or more optical lenses (in the figures, only one lens is used to represent a lens group for the sake of simplicity). It should be noted that if an optical lens group consists of only one optical lens, this optical lens is also referred to as an "optical lens group" in this application. These optical lenses are combined together such that the focal length of the entire optical lens group is f. The specific types of each optical lens are not limited. For example, it can be a convex lens as shown in Figure 3 , or it can be a curved mirror as shown in Figure 7 . The number of lenses included in each optical lens group is also not limited. For the sake of simplicity, in this application, each optical lens group only includes one optical lens. However, in practice, an optical lens group with a focal length of f can also be composed of multiple lenses. The specific method is well-known to those skilled in the art and will not be elaborated in this application.

[0062] In this application, the first phase plate is located at the focal points of the first optical lens group and the second optical lens group. The first phase plate can modulate light to achieve the first convolution operation, so that the first 4f optical system can, according to the first convolution operation, when the input is the target object, make the light intensity of the light signal output in the target area be the first light intensity. Among them, the target area can be understood as a part of the output area of the first 4f optical system (which can be understood as a part of the "image plane"), and the light intensity of the image in this part of the area is the first light intensity.

[0063] Optionally, the first light intensity can be a light intensity that has a positive or negative correlation with the light intensity of the ambient light, that is, greater than the light intensity of the ambient light (it can be greater than a difference or a ratio) or less than the light intensity of the ambient light (it can be less than a difference or a ratio). Correspondingly, in this application, the second light signal is used to eliminate the influence of the environment on the first light signal. Therefore, the second light intensity of the second light signal can also be related to the ambient light. Specifically, the second light intensity can have the opposite characteristic to the first light intensity. That is, when the input object includes the target object, when the first light intensity is greater than the light intensity of the ambient light, it can be recognized most accurately, and the greater the light intensity; at the same time, when the second light intensity is less than the light intensity of the ambient light, it can be recognized most accurately, and the smaller the light intensity; or when the first light intensity is less than the light intensity of the ambient light, the second light intensity is greater than the light intensity of the ambient light. Of course, the second light intensity can also be the light intensity of the ambient light or also have the opposite characteristic to the first light intensity (but still ensure a certain difference from the first light intensity). In this way, the influence of the ambient light on the first light signal can also be eliminated by comparing with the first light intensity, so as to output the final result by comparing the first light intensity and the second light intensity (specifically, by comparing the first electrical signal and the second electrical signal).

[0064] In this application, the first phase plate can be set such that each part has a corresponding thickness according to the requirements of the first convolution operation, so as to achieve the first convolution operation function through phase modulation. It can also change the light transmittance by covering a film that meets the requirements of the first convolution operation on each part, or have both characteristics, so as to achieve modulation by changing the light intensity and achieve the first convolution operation function. To achieve better results, the phase plate can also have the functions of both modulation methods simultaneously.

[0065] In another embodiment, the first phase plate can have multiple partitions, and each partition is used to perform a convolution operation on a specific target object (such as the faces of different ethnic groups). Among them, each partition has at least one of the following two characteristics: each part in the partition has a thickness that conforms to the convolution operation corresponding to that part of the partition, and each part in the partition has a light transmittance that conforms to the first convolution operation.

[0066] Each partition can identify a specific type of target object. For example, when the target object is a human face, different types of target objects can refer to human faces of different ethnic groups, such as the faces of Caucasians, Asians, Blacks, etc. Each partition is used to identify a specific type of target object, which can enable the first 4f optical system to output a first optical signal as long as any one partition recognizes a human face.

[0067] During training, each partition can be trained separately for a similar type of sample to obtain the final phase plate. The advantage of doing this is that during this training method, the similarity of the samples is higher, so the accuracy of the training will also be higher. Therefore, compared with the training method that includes multiple types of target objects (such as not distinguishing the faces of different ethnic groups) trained together, the final recognition rate is higher.

[0068] Embodiment III

[0069] Based on the above embodiments, this embodiment specifically introduces the reference optical system. In this embodiment, similar to the first 4f optical system in Embodiment II, the reference optical system in this embodiment is also a 4f optical system and has the second convolution operation ability, which is called the "second 4f optical system" in this application. Specifically, see Figure 3 , the second 4f optical system 22 includes: a third optical lens group 221, a second phase plate 222, and a fourth optical lens group 223 arranged in sequence along the optical path;

[0070] The third optical lens group 221 includes one or more third optical lenses, and the focal length of the third optical lens group is f;

[0071] The fourth optical lens group 223 includes one or more fourth optical lenses, and the focal length of the fourth optical lens group is f;

[0072] In this application, the second phase plate is located at the focal points of the third optical lens group and the fourth optical lens group. Similar to the first phase plate, the second phase plate may also have multiple thickness gradients obtained through AI training that can perform the second convolution operation and / or multiple films with different light transmittances, so that the second 4f optical system realizes that when the input is the target object, the light intensity of the second optical signal output in the target area is the second light intensity according to the second convolution operation.

[0073] The structure of the second 4f optical system is similar to that of the first 4f optical system. For the specific implementation of each component, reference can be made to the description of the first 4f optical system in Embodiment 2, which will not be elaborated in this embodiment. Different from Embodiment 2, the second phase plate is used to complete the second convolution operation, so that the output light intensity is the second light intensity. This second light intensity can be subsequently used to compare with the first light intensity, so that the final object recognition result can be output according to the comparison.

[0074] It can be understood that in order to make the magnitudes of the finally output first electrical signal (converted from the first light intensity) and the second electrical signal (converted from the second light intensity) more easily satisfy a certain difference relationship (such as the relative difference is greater than or equal to a certain threshold), the second light intensity can be opposite to the characteristics of the first light intensity. That is, when the input object includes the target object, when the first light intensity is greater than the ambient light intensity, it may be the most accurately recognized, and the greater the light intensity; at the same time, when the second light intensity is less than the ambient light intensity, it may be the most accurately recognized, and the smaller the light intensity. In this way, even if one of the 4f optical systems is not very accurately recognized, as long as the other 4f optical system can accurately recognize, then the magnitudes of the finally output first electrical signal and the second electrical signal will satisfy a certain difference relationship, so as to obtain the correct recognition result.

[0075] See Figure 4 , in another embodiment, the reference optical system includes one or more light intensity averaging devices arranged in sequence along the optical path. Among them, the light intensity averaging device is used to average the light intensity to prevent some points from being too bright or too dark. The light intensity averaging device is used to average the ambient light and then output the averaged ambient light to the second photosensitive detector.

[0076] The light intensity averaging device is also called a "rotating diffuser", and can be implemented by various methods. For example, simply by using frosted glass (also called matte glass, anti-glare glass, sandblasted glass, etc., which is a semi-transparent glass with a rough and uneven surface ground with emery or other substances or treated chemically).

[0077] In this embodiment, by averaging the light intensity of the ambient light, the second photosensitive detector can detect the ambient light more accurately, preventing errors in the detection results caused by over-bright or over-dark conditions in each area, thereby affecting the final judgment. In this embodiment, when the second photosensitive detector is used to detect the light intensity of the ambient light, the first light intensity of the first optical signal output by the first 4f optical system can be greater than or less than the light intensity of the ambient light. The two optical signals are respectively converted into corresponding electrical signals after passing through the photosensitive detectors 23 and 24, and then the comparator 25 is used to compare whether the relative difference between the two satisfies a specific threshold (for example, whether it is greater than or equal to the threshold). If it is satisfied, the recognition result is a successful recognition (that is, the target object is recognized); otherwise, the recognition result is an unsuccessful recognition (that is, the target object is not recognized).

[0078] See Figure 5 , in another embodiment, in order to further simplify the system, the reference optical system does not use any optical devices, so that the input ambient light directly passes through the reference optical system and is detected by the second photosensitive detector, thus achieving a simpler implementation.

[0079] Embodiment 4

[0080] Based on the above embodiments, this embodiment introduces a specific implementation of the first 4f optical system and the second 4f optical system in this application. It should be noted that for different modulation devices, there are different methods to train and manufacture the modulation device. For example, the convolution operation can be achieved by changing the thickness and / or light transmittance of each part of the phase plate (for example, a rectangular area from a few micrometers to dozens of micrometers). The following takes the modulation device as a phase plate as an example for illustration.

[0081] 1) The characteristics of the phase plate can be described by the following formula

[0082]

[0083] where is the transmittance at the coordinate , is the phase at the coordinate (related to the optical path). Its Fourier transform is denoted as

[0084]

[0085] where is the average value of the wavelength range that the photosensitive detector can detect, and f is the focal length of the lens.

[0086] The 4f system designed in this application performs the following convolution on the input image, and the corresponding convolution kernel is the absolute value square of the Fourier transform of the phase function of the phase plate

[0087]

[0088] where I in (x, y) is the light intensity at the input image (x, y), and I out (x, y) is the light intensity at the output image (x, y).

[0089] 2) By designing a specific transmittance distribution A and phase distribution φ, the required convolution operation can be performed on the image, that is, the obtained I out is the inversion of the result after convolution. The photosensitive detector detects the I out local light intensity:

[0090]

[0091] where "local" here refers to the range of the output image covered by the photosensitive detector, which depends on the position where the photosensitive detector is placed. It is required here that the photosensitive detector can cover a part of the output image area but not the entire output image area. The part of the output image area covered by the photosensitive detector is called the covered part, and the part of the output image area not covered by the photosensitive detector is called the non-covered part.

[0092] 3) In the above two 4f optical systems, different phase plates (phase plate 1 / phase plate 2) are designed respectively, that is, different transmittance distributions A 1 and A 2 ; different phase distributions φ 1 and φ 2 . Here, how to obtain the transmittance and phase distributions through AI training is introduced.

[0093] First, select a number of frontal face photos and non-frontal face photos as input images. Perform convolution operations on the input images

[0094]

[0095]

[0096] where is the light intensity distribution of the output image corresponding to phase plate 1, is the light intensity distribution of the output image corresponding to phase plate 2. Here, we set the maximum light intensity to 1 and the minimum light intensity to 0. Note that the same input image is obtained for both systems. The parameters of the convolution kernels K 1 and K 2 are the target parameters for training.

[0097] The definition of the cost function is

[0098]

[0099]

[0100] The goal of training is naturally to adjust the parameters of the convolution kernel to minimize the above cost function.

[0101] 4) Specifically, to train the convolution kernel parameters, set:

[0102] When the input image is a face

[0103] where (x, y) belongs to the covered part;

[0104] where (x, y) belongs to the non-covered part;

[0105] where (x, y) belongs to the covered part;

[0106] where (x, y) belongs to the non-covered part.

[0107] When the input image is not a face

[0108] where (x, y) belongs to the covered part;

[0109] where (x, y) belongs to the non-covered part;

[0110] where (x, y) belongs to the covered part;

[0111] where (x, y) belongs to the non-covered part.

[0112] The specific training steps are as follows:

[0113] First, assign a set of random initial values to the convolution kernel K 1 (x, y), K 2 (x, y) and obtain the initial and respectively to obtain the initial where,

[0114]

[0115]

[0116] Then, use the gradient descent method to continuously adjust K 1 (x, y) and K 2 (x, y) to make the cost function L 1 , L2 Continuously shrink and finally obtain a target K that can make the cost function L 1 , L 2 meet the requirements (such as being minimized). 1 (x, y) and the target K 2 (x, y). Specifically, the transmittance and the phase function can be adjusted to adjust K 1 (x, y) and K 2 (x, y).

[0117] Finally, obtain the target convolution kernel function target K 1 (x, y) and the target K 2 (x, y). Since the target K 1 (x, y) 1 and the target K 2 (x, y) are adjusted by the transmittance and the phase function , therefore, when obtaining the target K 1 (x, y) and the target K 2 (x, y), the transmittance and the phase function can also be obtained simultaneously (as long as the values of these two parameters are recorded when adjusting the target convolution kernel function using the transmittance and the phase function ).

[0118] The phase function can be realized by controlling the phase plate thickness function height_map . By controlling the local thickness of each coordinate of the phase plate, different optical path differences can be formed, thereby realizing different phase differences. Therefore, the target phase plate thickness function height_map , that is, the thickness value at each coordinate , can be obtained from the following equation:

[0119]

[0120] where represents the target phase value corresponding to the coordinate , λ represents the wavelength of the system, n represents the refractive index of the material used for the phase plate, and Δn represents the difference between the refractive index of the phase plate and the refractive index of air. In order to achieve the target thickness function height_map , it can be fabricated as follows: First, layer by layer according to the gradient, draw the mask of this layer; then perform multiple etching operations layer by layer to obtain a phase plate with multiple thickness gradients.

[0121] Transmittance It can be achieved by further coating the phase plate obtained in the above steps: uniformly deposit the oxide on the surface of the glass sheet, and use different oxide film groups to precisely control the transmittance of light with corresponding wavelengths. It should be noted that the specific process for manufacturing the phase plate in this application is prior art and will not be elaborated in this application.

[0122] If a phase plate includes each partition described in the second embodiment, each partition can be trained separately according to the above method, so that as long as one partition meets the requirements finally, the phase plate can output the corresponding optical signal.

[0123] See Figure 6 , which is a schematic structural diagram of a manufactured phase plate.

[0124] The manufactured phase plate has the following properties:

[0125] When the input object includes a human face:

[0126] After passing through the upper 4f optical system (the system corresponding to phase plate 1), the light intensity in the detection area of photosensitive detector 1 is high;

[0127] After passing through the lower 4f optical system (the system corresponding to phase plate 2), the light intensity in the detection area of photosensitive detector 2 is low.

[0128] When the input object does not include a human face:

[0129] After passing through the upper 4f optical system (the system corresponding to phase plate 1), the light intensity in the detection area of photosensitive detector 1 is low;

[0130] After passing through the lower 4f optical system (the system corresponding to phase plate 2), the light intensity in the detection area of photosensitive detector 2 is high.

[0131] Embodiment Five

[0132] Based on the above embodiments, this embodiment introduces another optical path implementation form of the device of this application.

[0133] It can be understood that in this application, the 4f optical system of the foregoing embodiments can be completed through various specific optical path settings. For example, as Figure 2 shown in the optical path based on a straight line, or an optical path based on a non - straight line can also be adopted. For example, as Figure 7 shown, use a curved mirror to replace the convex lens in the foregoing embodiments, thereby changing the optical path.

[0134] Specifically, Figure 3The optical device includes a first 4f optical system 51, a second 4f optical system 52, a first photosensitive detector 53, a second photosensitive detector 54, and a comparator 55. The first 4f optical system 51 includes a first curved mirror 511, a first phase plate 512, and a second curved mirror 513. For the optical path of the first 4f optical system 51, reference can be made to Figure 3 the direction indicated by the arrow in

[0135] i.e., the incident light is first reflected by 511 to the first phase plate 512, and after passing through the first phase plate, it is then reflected by the second curved mirror 513 to the first photosensitive detector 53.

[0136] It can be understood that in addition to the optical paths in this embodiment, there can be more optical paths to implement the 4f optical system by using principles such as refraction and reflection of light, which is not limited in this application. In practical applications, appropriate optical paths can be selected according to the hardware structures of different products.

[0137] Embodiment Six

[0138] Based on the above embodiments, this embodiment specifically introduces the design of the light inlet. As can be seen from the above embodiments, since there is a first 4f optical system and a reference optical system, both of these systems need to receive light. When designing the light inlet, one method is to allow external light to enter both systems simultaneously. For example, as shown in Figure 8 (a) in

[0139] there are two light inlets (61-1 and 61-2), and external light can enter both systems through these two light inlets simultaneously. Figure 8 Referring to

[0140] Figure (b) in

[0141] in another implementation, it can also be designed with only one light inlet 61, and then the incident light is transmitted to each system through optical devices. For example, the external light can be received through the light inlet 61 first, and then a part of the light is transmitted to one system (such as the reference optical system in the figure) through a beam splitter 62 (for example, a semi-transparent and semi-reflective mirror), and the other part of the light is propagated in another direction through the reflection of the semi-reflective mirror 62; then, the light propagated in the other direction is introduced into another system (such as the first 4f optical system shown in the figure) through another optical device (such as the mirror 63 shown in the figure). Of course, in practice, other optical devices can also be used to perform various optical path changes such as reflection or transmission of light to form different optical paths to adapt to different products.

[0140] Embodiment Seven

[0141] Based on the above embodiments, this embodiment discloses various components for object recognition. It can be understood that in the actual application process, one or more of the various components in the object recognition device in the above embodiments can be encapsulated into a component, so that it is more convenient to use in the process of use.

[0142] For example, in one embodiment, all components (including all 4f optical systems, all photosensitive detectors, and comparators) can be encapsulated into a component.

[0143] In another embodiment, all 4f optical systems and all photosensitive detectors can also be encapsulated into a component, without encapsulating the comparator.

[0144] In another embodiment, two 4f optical systems can also be encapsulated into a component, or the two 4f optical systems can be encapsulated into a component separately.

[0145] In another embodiment, one 4f optical system and the photosensitive detector corresponding to this 4f optical system can be encapsulated into a component, and another 4f optical system and the photosensitive detector corresponding to this 4f optical system can be encapsulated into a component.

[0146] The above are several specific examples. In actual applications, there is no limitation to encapsulate the various components involved in this application based on other combinations (such as combinations of any two or three), and of course, encapsulation can also not be performed.

[0147] Embodiment Eight

[0148] Based on the above embodiments, refer to Figure 9 , this application discloses an electronic device 80. This electronic device can be a terminal device (such as a mobile phone, a tablet computer, a personal computer, etc.), or can also be various communication devices for enterprises (such as base stations, servers, switches, routers, etc.). The electronic device includes a processing circuit 81 and the object recognition device 82 in the foregoing embodiments, wherein the processing circuit 81 is used to perform processing based on the processing result of the object recognition device.

[0149] The processing circuit 81 may be a software-implemented processing circuit or a hardware-implemented processing circuit. In this application, "software-implemented" means that a processor reads and executes program instructions stored in a memory to implement the functions corresponding to the above-mentioned modules or units. Here, the processor refers to a processing circuit having the function of executing program instructions, including but not limited to at least one of the following: central processing unit (CPU), microprocessor, digital signal processor (DSP), microcontroller unit (MCU), or various processing circuits such as artificial intelligence processors that can run program instructions. In some other embodiments, the processor may further include circuits with other processing functions (such as hardware circuits for hardware acceleration, bus and interface circuits, etc.). The processor may be presented in the form of an integrated chip. For example, it may be presented in the form of an integrated chip whose processing function only includes the function of executing software instructions, or it may also be presented in the form of a system on a chip (SoC). That is, on one chip, in addition to including a processing circuit that can run program instructions (usually referred to as a "core"), it also includes other hardware circuits for implementing specific functions (of course, these hardware circuits can also be separately implemented based on ASIC or FPGA). Correspondingly, in addition to the function of executing software instructions, the processing function may also include various hardware acceleration functions (such as AI computing, encoding and decoding, compression and decompression, etc.).

[0150] For example, in one example, if the electronic device is a mobile phone, the processing circuit can be the processor in the mobile phone (such as the Kirin series processors of Huawei or the Snapdragon series processors of Qualcomm). To store the code of the programs required for the processor to run, the electronic device further includes a memory 83. The processor can run software such as an operating system and application programs, and perform various processes based on the output result of the object recognition device 82. For example, a typical application scenario is to install the object recognition device on the front of the mobile phone (similar to the position where the front camera is installed) and use it to identify whether there is a human face looking at the screen (for example, when the user may be reading the content on the mobile phone). If the object recognition device recognizes a human face, it will output a relevant signal to the processor through a comparator. The processor will keep the screen from dimming or turn off the screen according to this signal. In this way, the screen will always be on when the user is reading, providing a better reading experience. If no human face is recognized, the screen can be turned off or the brightness can be reduced, thereby reducing power consumption. In another application scenario, user privacy can also be protected based on object recognition. For example, if multiple human faces are recognized, some sensitive information on the screen (such as previewed text messages or information from social software) can be not displayed. Or, if sensitive information is being read at that time, the screen can be dimmed or turned off to prevent the information from being seen by others. Of course, in this scenario, multiple human faces need to be recognized. Therefore, during training, those skilled in the art can make adaptive adjustments to the training target based on the above method so that the device can recognize multiple human faces.

[0151] In this application, "implemented based on hardware" means that the functions of the above-mentioned modules or units are implemented through a hardware processing circuit that does not have the function of processing program instructions. This hardware processing circuit can be composed of discrete hardware components or can be an integrated circuit. To reduce power consumption and size, the integrated circuit form is usually adopted for implementation. The hardware processing circuit can include an ASIC (application-specific integrated circuit), or a PLD (programmable logic device); among them, the PLD can further include an FPGA (field programmable gate array), a CPLD (complex programmable logic device), and so on. These hardware processing circuits can be a separately packaged semiconductor chip. For example, multiple hardware circuits and a CPU can be formed on a silicon substrate and separately packaged into a chip, and this chip is also called an SoC. Or, a circuit for implementing the FPGA function and a CPU can be formed on a silicon substrate and separately enclosed into a chip, and this chip is also called an SoPC (system on a programmable chip).

[0152] It should be noted that the processing circuit in this application is not limited to being implemented in a software-hardware combination manner. For example, a part of the processing circuit is implemented using a CPU, and another part is implemented using an FPGA. Therefore, those skilled in the art can choose to implement in the form of software, hardware, or a combination of both according to actual needs.

[0153] Those of ordinary skill in the art can understand that all or part of the processes of implementing the methods in the above embodiments can be completed by instructing relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, the storage medium can be a magnetic disk, an optical disc, a read-only memory (ROM), or a random access memory (RAM), etc.

[0154] The above-mentioned preferred embodiments have further detailed the purpose, technical solutions, and advantages of the present invention. It should be understood that the above is only the preferred embodiments of the present invention and is not used to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An object recognition device, characterized in that, it includes a first 4f optical system, a reference optical system, a first photosensitive detector, a second photosensitive detector, and a comparator; the first 4f optical system is configured to output a first optical signal with a first light intensity according to a first input object, wherein the first input object includes a target object; the first photosensitive detector is configured to convert the first optical signal into a first electrical signal; the reference optical system is configured to output a second optical signal with a second light intensity according to the first input object, the first light intensity and the second light intensity are related to ambient light, and the second light intensity has a characteristic opposite to that of the first light intensity; the second photosensitive detector is configured to convert the second optical signal into a second electrical signal; the comparator is configured to compare the first electrical signal and the second electrical signal, and output an identification result indicating that the target object has been identified according to the comparison result between the first electrical signal and the second electrical signal.

2. The device according to claim 1, characterized in that, the relative difference between the magnitude of the first electrical signal and the magnitude of the second electrical signal is greater than or equal to a first preset threshold.

3. The device according to claim 1 or 2, characterized in that: the first 4f optical system is further configured to output a third optical signal with a third light intensity according to a second input object, wherein the second input object does not include the target object; the first photosensitive detector is further configured to convert the third optical signal into a third electrical signal; the reference optical system is further configured to output a fourth optical signal with a fourth light intensity according to the second input object; the fourth light intensity is not equal to the third light intensity; the second photosensitive detector is further configured to convert the fourth optical signal into a fourth electrical signal; the comparator is further configured to compare the third electrical signal and the fourth electrical signal, and output an identification result indicating that the target object has not been identified according to the comparison result between the third electrical signal and the fourth electrical signal.

4. The device according to claim 3, characterized in that, the relative difference between the magnitude of the third electrical signal and the magnitude of the fourth electrical signal is less than a second preset threshold.

5. The device according to claim 1 or 2, characterized in that, the first 4f optical system is a 4f optical system with a first convolution operation ability, and the first 4f optical system is configured to output the first optical signal based on the first convolution operation.

6. The device according to claim 5, characterized in that, the first 4f optical system includes: a first optical lens group, a first phase plate, and a second optical lens group arranged in sequence along the optical path; the first optical lens group includes one or more first optical lenses, and the focal length of the first optical lens group is f; the second optical lens group includes one or more second optical lenses, and the focal length of the second optical lens group is f; the distance between the first phase plate and the first optical lens group and the distance between the first phase plate and the second optical lens group are both f; the first phase plate is configured to perform the first convolution operation.

7. The device according to claim 6, characterized in that: the first phase plate has at least one of the following two characteristics: Each part of the first phase plate has a thickness conforming to the first convolution operation, and each part of the first phase plate has a light transmittance conforming to the first convolution operation.

8. The device according to claim 1 or 2, characterized in that the first 4f optical system includes: a first optical lens group, a first phase plate, and a second optical lens group arranged in sequence along the optical path; the first optical lens group includes one or more first optical lenses, and the focal length of the first optical lens group is f; the second optical lens group includes one or more second optical lenses, and the focal length of the second optical lens group is f; the distances between the first phase plate and the first optical lens group and between the first phase plate and the second optical lens group are both f; the first phase plate has a plurality of partitions, and each partition is used for performing a convolution operation on a specific target object. Among them, each partition has at least one of the following two characteristics: each part in the partition has a thickness conforming to the convolution operation corresponding to the partition, and each part in the partition has a light transmittance conforming to the convolution operation corresponding to the partition.

9. The device according to claim 1 or 2, characterized in that the reference optical system includes a second 4f optical system having a second convolution operation ability, and the second 4f optical system is used to output the second optical signal based on the second convolution operation.

10. The device according to claim 9, characterized in that the second 4f optical system includes: a third optical lens group, a second phase plate, and a fourth optical lens group arranged in sequence along the optical path; the third optical lens group includes one or more third optical lenses, and the focal length of the third optical lens group is f; the fourth optical lens group includes one or more fourth optical lenses, and the focal length of the fourth optical lens group is f; the distances between the second phase plate and the third optical lens group and between the second phase plate and the fourth optical lens group are both f; the second phase plate is used to perform the second convolution operation.

11. The device according to claim 10, characterized in that: the second phase plate has at least one of the following two characteristics: each part of the second phase plate has a thickness conforming to the second convolution operation, and each part of the second phase plate has a light transmittance conforming to the second convolution operation.

12. The device according to claim 9, characterized in that the second 4f optical system includes: a third optical lens group, a second phase plate, and a fourth optical lens group arranged in sequence along the optical path; the third optical lens group includes one or more third optical lenses, and the focal length of the third optical lens group is f; the fourth optical lens group includes one or more fourth optical lenses, and the focal length of the fourth optical lens group is f; the distances between the second phase plate and the third optical lens group and between the second phase plate and the fourth optical lens group are both f; the second phase plate has a plurality of partitions, and each partition is used for performing a convolution operation on a specific target object. Among them, each partition has at least one of the following two characteristics: each part in the partition has a thickness conforming to the convolution operation corresponding to the partition, and Each part in the partition has a light transmittance that conforms to the convolution operation corresponding to the partition.

13. The device according to claim 1 or 2, wherein, the reference optical system includes a light intensity averaging device; the light intensity averaging device is used to average the intensity of the ambient light and then output it to the second photosensitive detector.

14. The device according to claim 1 or 2, wherein, the device further includes a beam splitter, and the beam splitter is used to split the input optical signal into two paths and output them to the first 4f optical system and the reference optical system respectively.

15. An electronic device, wherein, it includes a processing circuit and the object recognition device according to any one of claims 1 - 14, wherein the processing circuit is used to process based on the recognition result output by the object recognition device.

16. The electronic device according to claim 15, wherein, the electronic device further includes a display screen, and when the processing circuit is used to process based on the recognition result output by the object recognition device, it is specifically used to determine whether to turn off the display of the display screen based on the recognition result.

17. An optical component, wherein, the optical component includes a first 4f optical system, a reference optical system, a first photosensitive detector, and a second photosensitive detector; the first 4f optical system is used to output a first optical signal with a first light intensity according to a first input object, and the first input object includes a target object; the first photosensitive detector is used to convert the first optical signal into a first electrical signal; the reference optical system is used to output a second optical signal with a second light intensity according to the first input object; the first light intensity and the second light intensity are related to the ambient light, and the second light intensity has the opposite characteristic to the first light intensity; the second photosensitive detector is used to convert the second optical signal into a second electrical signal; wherein, when the first electrical signal and the second electrical signal are compared, the comparison result obtained through the comparison is used to determine that the first input object is recognized to contain the target object.

18. The optical component according to claim 17, wherein, the first 4f optical system is further used to output a third optical signal with a third light intensity according to a second input object, where the second input object does not include the target object; the first photosensitive detector is further used to convert the third optical signal into a third electrical signal; the reference optical system is further used to output a fourth optical signal with a fourth light intensity according to the input object; the fourth light intensity is not equal to the third light intensity; the second photosensitive detector is further used to convert the fourth optical signal into a fourth electrical signal; wherein, when the third electrical signal and the fourth electrical signal are compared, the comparison result obtained through the comparison is used to determine that the second input object is recognized not to contain the target object.

19. The optical component according to claim 17 or 18, wherein, the first 4f optical system is a 4f optical system with a first convolution operation ability, and the first 4f optical system is used to output the first optical signal based on the first convolution operation. Among them, the first 4f optical system includes: A first optical lens group, a first phase plate, and a second optical lens group arranged in sequence along the optical path; The first optical lens group includes one or more first optical lenses, and the focal length of the first optical lens group is f; The second optical lens group includes one or more second optical lenses, and the focal length of the second optical lens group is f; The distance between the first phase plate and the first optical lens group and the distance between the first phase plate and the second optical lens group are both f; The first phase plate is used to perform the first convolution operation; The first phase plate has at least one of the following two characteristics: Each part of the first phase plate has a thickness conforming to the first convolution operation, and Each part of the first phase plate has a light transmittance conforming to the first convolution operation.

20. The optical component according to claim 17 or 18, characterized in that The first 4f optical system includes: A first optical lens group, a first phase plate, and a second optical lens group arranged in sequence along the optical path; The first optical lens group includes one or more first optical lenses, and the focal length of the first optical lens group is f; The second optical lens group includes one or more second optical lenses, and the focal length of the second optical lens group is f; The distance between the first phase plate and the first optical lens group and the distance between the first phase plate and the second optical lens group are both f; The first phase plate has a plurality of partitions, and each partition is used to perform a convolution operation on a specific target object. Among them, each partition has at least one of the following two characteristics: Each part in the partition has a thickness conforming to the convolution operation corresponding to the partition, and Each part in the partition has a light transmittance conforming to the convolution operation corresponding to the partition.

21. The optical component according to claim 17 or 18, characterized in that The reference optical system includes a second 4f optical system with the ability of a second convolution operation, and the second 4f optical system is used to output the second optical signal based on the second convolution operation; the second 4f optical system includes: A third optical lens group, a second phase plate, and a fourth optical lens group arranged in sequence along the optical path; The third optical lens group includes one or more third optical lenses, and the focal length of the third optical lens group is f; The fourth optical lens group includes one or more fourth optical lenses, and the focal length of the fourth optical lens group is f; The distance between the second phase plate and the third optical lens group and the distance between the second phase plate and the fourth optical lens group are both f; The second phase plate is used to perform the second convolution operation; The second phase plate has at least one of the following two characteristics: Each part of the second phase plate has a thickness conforming to the second convolution operation, and Each part of the second phase plate has a light transmittance conforming to the second convolution operation.

22. The optical component according to claim 17 or 18, characterized in that The reference optical system includes a second 4f optical system with the ability of a second convolution operation, and the second 4f optical system is used to output the second optical signal based on the second convolution operation; the second 4f optical system includes: The third optical lens group, the second phase plate, and the fourth optical lens group arranged in sequence along the optical path; The third optical lens group includes one or more third optical lenses, and the focal length of the third optical lens group is f; The fourth optical lens group includes one or more fourth optical lenses, and the focal length of the fourth optical lens group is f; The distances between the second phase plate and the third optical lens group and between the second phase plate and the fourth optical lens group are both f; The second phase plate has a plurality of partitions, and each partition is used for performing a convolution operation on a specific target object. Among them, each partition has at least one of the following two characteristics: Each part in the partition has a thickness that conforms to the convolution operation corresponding to the partition, and Each part in the partition has a light transmittance that conforms to the convolution operation corresponding to the partition.

23. The optical component according to claim 17 or 18, characterized in that The optical component further includes a beam splitter, and the beam splitter is used to split the input optical signal into two paths and output them to the first 4f optical system and the reference optical system respectively.

Images