System and method for high-resolution imaging of an object using an event camera

Abstract

The invention relates to imaging of an object to generate a high-resolution image IMA of the object with the help of an imaging system comprising a DVS camera and a beam shifting device. Imaging data are collected with a sensor chip of the camera while a relative lateral shift of a projection P200 of the object and the sensor chip is performed. From the 1 resulting imaging data, the development of light flux over time onto a given sensor element of the chip is reconstructed. The so determined light flux over time is evaluated to determine the position of a pixel of the high-resolution image as well as the pixel value to be assigned to that pixel.

Description

[0001] Description

[0002] SYSTEM AND METHOD FOR HIGH-RESOLUTION IMAGING OF AN OBJECT USING AN EVENT CAMERA

[0003] The invention relates to imaging of an obj ect to generate a high-resolution image IMA of the obj ect .

[0004] Optical imaging e . g . in mass production suf fers the problem of a huge amount of data being created when obj ects shall be inspected with high lateral resolution, e . g . at their velocity of transport in the production . Typical data rates caused by corresponding camera systems might already nowadays be in the range above 1 or even 5 Gbit / s . This amount of data needs to be transmitted and pre-evaluated in a smart camera or an edge device in order to make the data stream to an end device like a superordinated production control system manageable . Lower data rates would result in less resolution and, therewith, lower imaging accuracy which would have a negative impact on the quality of the production process .

[0005] Apart from such moving obj ect imaging tasks , any optical precision measurement always requires high resolution at a respective obj ect site of interest . In the light of the data volume to remain manageable while , at the same time , the spatial and / or temporal resolution of the imaging task shall be high, one concrete problem to be solved might be how to achieve higher resolution without j ust ramping up the overall pixel number of the imaging system and thus blowing up the amount of generated data .

[0006] State of the art imaging systems capture the image data in high volume with a suitable camera and transmit captured data to an edge device which performs data pre-evaluation or even full evaluation . In alternative solutions , the data is transmitted to a cloud system for analysis . In some cases , smart cameras are used where CPUs or FPGAs are incorporated in order to identi fy defects at an early stage of the process and to crop a relevant area of interest in the images or to aggregate the data e.g. by packaging etc. In many applications line-scan cameras are in use in order to have a more constant data rate than with full field cameras, but the general topic of the high data volume stays exactly the same for the same resolution. Such imaging technologies provide the required high spatial resolution. However, they do not include solutions which, on the one side, provide images with low data volumes and which, on the other side, cover large dynamic ranges of e.g. 80 to lOOdB or even higher. The high dynamic range is necessary to cover different illumination conditions in the field of view of the camera and to maintain good imaging properties to ensure the high imaging accuracy and quality. Typical full field or line cameras only have 40 to 60dB of maximal dynamic range. However, the so called "event" or "DVS" cameras, e.g. described in PCT / EP2022 / 070785 and, for 3D applications, in PCT / EP2022 / 074649, promise to be an imaging technology with the potential to overcome those issues of data volume, dynamic range, and spatial resolution.

[0007] Therefore, a solution is required which allows fast and precise high-resolution imaging with low data volume. This is solved by the method suggested in claim 1 and by the imaging system as per claim 14. Dependent claims introduce advantageous embodiments.

[0008] A Method METH for camera based imaging of an object for generating a high-resolution image IMA of an object with subpixel accuracy utilizes an imaging system, which comprises at least the camera itself and a beam shifting device. The camera is embodied as a DVS camera with a light sensitive sensor chip with a plurality of sensor elements SE(i) , each of which generates imaging data DAT(i) only in case of an event EV(i) of change of light flux EX onto that sensor element SE(i) by more than a given threshold THRES, e.g. THRES=15%. Thus, such a circumstance of the amount of light flux change being higher than the threshold THRES is called an "event" EV(i) . Generated imaging data DAT(i) include a polarity POL(i) representing the sign of change of light flux, i.e. positive POS or negative NEG, a time stamp TS (i) , and an identifier ID(i) of the imaging data DAT(i) generating sensor element SE(i) . The beam shifting device is configured to effect a relative lateral shift of the sensor chip and a projection P200 of at least a section of the object onto the sensor chip, i.e. the projection P200 and / or the sensor chip itself might be shifted to effect the relative lateral shifting. The projection P200 is a visual representation of at least a section of the object the sensor chip. An optics section might be foreseen for converting a light flux from the object into the projection P200.

[0009] The method METH comprises a stage INSPECT of observing the object with the DVS camera to generate the imaging data DAT(i) , including execution of one or more continuous or discrete shifting sequences SEQ(k) . During each one of the shifting sequences SEQ(k) , i.e. over time t, a relative lateral shifting of the projection P200 and the sensor chip is executed by means of the beam shifting device, such that the projection P200 and the sensor chip (111) are shifted relative to each other in a known shifting manner MAN, i.e. with known temporal shifting behavior, i.e. speed, in a known direction, and by a known shifting amount or, in one word, with a known trajectory, to generate the imaging data DAT(i,k) . For each sensor element SE(i) , generated imaging data DAT(i,k) are collected for further processing. Moreover, the method METH comprises a stage EVAL of evaluating the generated imaging data DAT(i,k) for selected ones SE(il) of the sensor elements SE(i) . Therein, in the stage EVAL, for each selected sensor element SE(il) , a temporal light flux representation FXrec(SE ( il ) , t , k) is generated from the imaging data DAT(il,k) , representing the temporal development of light flux FX(t) caused by the projection P200 onto the respective selected sensor element SE(il) during the shifting sequence SEQ(k) , and pixel values PXVAL(j) of pixels PX(j) of the image IMA are determined based on the light flux representations FXrec(SE ( il ) , t , k) . In more concrete, the known shifting manner MAN includes, for the respective shifting sequence SEQ(k) , the shifting direction DIRSEQ(k) and information about the temporal behavior of shifting, e.g. the relative speed of the sensor chip and the projection P200. Thus, the knowledge of the shifting manner MAN enables, for each point in time during the shifting sequence SEQ(k) , determination how far and in which direction a certain, arbitrary point of the projection P200 has moved across the sensor chip. In turn, an original position of a certain point of the projection P200 before the shifting sequence SEQ(k) has started, can be reconstructed from the shifting manner MAN and the position POS of that point at a certain time T during the shifting sequence SEQ (k) .

[0010] Preferably, the overall shifting amount DSEQ(k) of a shifting sequence SEQ(k) is larger than the sensor chip's pixel pitch (dPXx, dPXy) in the respective shifting direction.

[0011] Especially in case the imaging data DAT(il,k) generated by the respective sensor element SE(il) during the shifting sequence SEQ(k) include a plurality of polarities POL(il,k,p) and a respective plurality of time stamps TS (il,k,p) with p=l,...,PLU and PLU representing the number of events EV(il,k) generated by sensor element SE(il) during the shifting sequence SEQ(k) , the light flux representation FXrec(SE ( il ) , t , k) is generated from the polarities POL(il,k,p) and respective time stamps TS (il,k,p) of the imaging data DAT(il,k) from an arbitrary starting value VALO, e.g. REF=10, by stepwise, i.e, for each p, starting from p=l, increase, preferably by a given factor FACT, preferably corresponding to the threshold THRES, e.g. FACT=THRES, in case of positive polarity POL(il,k,p) , i.e. VAL (n+1 ) =VAL (n) *FACT, and stepwise decrease, preferably by the given factor FACT, in case of negative polarity POL(il,k,p) , i.e. VAL (m+1 ) =VAL (m) / FACT . Due to the stepwise approach, the light flux representation FXrec(SE ( il ) , t , k) is not a continuous function, but it consists of a plurality of discrete data points. Preferably, the light flux representation FXrec(SE ( il ) , t , k) is further processed to determine the point Textrin time at which the light flux representation FXrec(SE ( il ) , t , k) or its derivative EX'rec(SE ( il ) , t , k) = dFXrec(SE ( il ) , t , k) / dt has an extreme value and

[0012] - the respective value EXT=FXrec(SE ( il ) , Textr, k) itself.

[0013] Therein, the identity of the pixel PX(jl) , i.e. the position of the respective pixel PX(jl) in the image IMA, for which a pixel value PXVAL(jl) is to be determined, is derived from the point Textrin time and from the known shifting manner MAN. The knowledge of the shifting manner MAN enables, for each point in time during the shifting sequence SEQ(k) , determination how far and in which direction a certain, arbitrary point of the projection P200 has moved relative to the sensor chip. In turn, an original position of a certain point of the projection P200 before the shifting sequence SEQ(k) has started, can be reconstructed from the shifting manner MAN and the position POS of that point at a certain time T during the shifting sequence SEQ(k) . Such "original position" corresponds to the identity of the pixel PX(jl) introduced above.

[0014] The value EXT can be used to determine the pixel value PXVAL(jl) of the pixel PX(jl) , preferably such that PXVAL ( j 1 ) =EXT . With that, grey values in the image IMA can be determined .

[0015] The further processing of the light flux representation

[0016] FXrec(SE ( il ) , t , k) might include fitting or interpolating the light flux representation FXrec(SE ( il ) , t , k) to determine an interpolated light flux representation FXrec, fit (SE ( il ) , t , k) . Therein, Textrshall be the point in time at which the interpolated light flux representation FXrec, fit (SE ( il ) , t , k) or its derivative EX'rec,fit(SE ( il ) , t , k) = dFXrec,fit(SE ( il ) , t , k) / dt has an extreme value and the respective value EXT shall be the value FXrec,fit(SE ( il ) , Textr, k) itself. In contrast to the original, discrete light flux representation

[0017] FXrec(SE ( il ) , t , k) , the interpolated light flux representation FXrec, fit (SE ( il ) , t , k) is a continuous function so that Textrcan be determined with higher accuracy.

[0018] Regarding the "selected" sensor elements, either all signal elements SE(i) can be selected signal elements SE(il) or only those ones SE(il) of the signal elements SE(i) are selected signal elements SE(il) , the imaging data DAT(il) over time of which correspond to a predefined pattern.

[0019] Therein the predefined pattern might be characterized by a time series of at last PLU>1 polarities POL(i,k) of the same kind, i.e. either POS or NEG, in a row. For example, K>20 might be reasonable.

[0020] The stage INSPECT might include a plurality of shifting sequences SEQ(k) with k=l,...,K and K>1, wherein different shifting sequences SEQ(kl) , SEQ(k2) with kl^k2 differ from each other with regard to the respective shifting directions DIRSEQ ( kl ) T^DIRSEQ ( k2 ) . This includes, for example, that the directions DIRSEQ(kl) , DIRSEQ(k2) are perpendicular on each other. However, it also includes that they are anti-parallel.

[0021] The lateral shifting during each shifting sequence SEQ(k) might be executed such that the projection P200 moves relative to the sensor chip along a straight line, preferably along columns or rows of the sensor elements SE(i) on the sensor chip. Thus, the shifting direction during one particular shifting sequence SEQ(k) is not changed.

[0022] The relative lateral shifting of the projection P200 and the sensor chip might be executed with the help of the beam shifting device by shifting the projection P200 relative to the sensor chip, preferably by controlling an optics section of the DVS camera, and / or shifting the sensor chip relative to the projection P200. The optics section of the DVS camera might comprise one or more integrated optical devices for generating the proj ection P200 of the obj ect from the light flux field LFL emitted by the obj ect , wherein the beam shi fting device might be reali zed by at least one of the optical devices of the optics section, wherein the lateral shi fting is achieved by changing optical properties of the at least one optical device , preferably by shi fting, tilting, and / or deforming the at least one optical device .

[0023] The imaging system for camera based imaging of the obj ect for generating the high-resolution image IMA of the obj ect with sub-pixel accuracy is configured to execute the method METH and comprises the camera, which is embodied as a DVS camera as described above , and the beam shi fting device for ef fecting the relative lateral shi ft of the sensor chip and the proj ection P200 . Moreover, the imaging system comprise control system with a control unit for controlling the beam shi fting device and the sensor chip to generate the imaging data DAT ( i ) and an evaluation unit for executing the stage EVAL for evaluating the imaging data DAT ( i ) to generate the high-resolution image IMA.

[0024] The camera of the imaging system might comprise an optics section for converting a light flux field LFL from the obj ect into the proj ection P200 of the obj ect onto the sensor chip, wherein the optics section comprises one or more integrated optical devices for generating the proj ection P200 from the light flux field LFL, wherein the beam shi fting device is reali zed by at least one of the optical devices of the optics section, wherein the lateral shi fting is achieved by changing electrical and / or optical properties of the at least one optical device , preferably by shi fting, tilting, and / or deforming the at least one optical device .

[0025] In more detail , in case the beam shi fting device is reali zed by at least one of the optical devices , the lateral shi fting can be achieved by changing the electrical and / or optical properties of the at least one optical device , preferably by shi fting, tilting, and / or deforming the at least one optical device ( 112 ’ ) . Therein, the at least one of the optical devices might comprise a pair of mirrors , might be embodied as a transparent , piano-parallel plate , and / or might be embodied as an electro-optical device , each of which are described in more detail below .

[0026] The approach introduced herein applies a DVS camera and achieves sub-pixel resolution by beneficial data processing . Thus , the special properties of a DVS camera as introduced below are beneficially applied to determine a physical feature FEAT of the obj ect with high resolution .

[0027] The invention advantageously implements a camera applying the "Dynamic Vision System" ( DVS ) principle , also known as "event camera" or "neuromorphic camera" . The essential properties of a DVS camera applied in the solution described herein include that the DVS camera detects events , i . e . changes of light flux EX or intensity, respectively, exceeding a pre-defined threshold value THRES of e . g . 15% , and that such change of light flux EX is detected pixelwise . In ef fect , the DVS camera has an ef fective dynamic range of 120db or even higher, as a result of the sensor elements ' full well capacity and adoptions of exposure time according to the minimum and maximum allowable values . The high dynamics of signals for an event on a sensor element being signi ficantly faster than a typical camera frame readout time and the high dynamic range of the camera and / or sensing principle are core aspects of the favorable solution provided herein . The signal rate of events can be in the range down to single microseconds while typical frame rates of Mpxl cameras are nowadays in the range of 10 to 5ms for 100 or 200Hz frame rates , respectively . Thus , the speed of the measuring system is increased by 3 to 4 orders of magnitude with respect to today' s technology . In summary, maj or di f ferences of the DVS camera based approach for obj ect imaging over conventional imaging systems are based on the properties of a DVS camera to detect events instead of full frame imaging, on the signi ficantly higher speed for DVS camera data acquisition compared to typical frame rates of full frame cameras , and on the signi ficant reduction in data volume , since only relevant events get reported instead of a full image capture of the total surface of the obj ect under inspection .

[0028] It is to be understood that the elements and features recited in the appended claims may be combined in di f ferent ways to produce new claims that likewise fall within the scope of the present invention . Thus , whereas the dependent claims appended below depend from speci fic independent or dependent claims , it is to be understood that these dependent claims can, alternatively, be made to depend in the alternative from any preceding or following claim, whether independent or dependent , and that such new combinations are to be understood as forming a part of the present speci fication .

[0029] DESCRIPTION OF THE FIGURES

[0030] In the following, possible embodiments of the di f ferent aspects of the present invention are described in more detail with reference to the enclosed figures . The obj ects as well as further advantages of the present embodiments will become more apparent and readily appreciated from the following description of the preferred embodiments , taken in conj unction with the accompanying figure in which :

[0031] FIG 1 shows an exemplary application scenario of an imaging system with a DVS camera,

[0032] FIG 2 shows the method METH for imaging,

[0033] FIG 3 shows a top view onto a sensor chip,

[0034] FIG 4 a representation of light flux onto a sensor element in case of a point like feature , FIG 5 a representation of light flux onto a sensor element in case of an edge like feature ,

[0035] FIG 6 shows a first embodiment of a beam shi fting device ,

[0036] FIG 7 shows a second embodiment of a beam shi fting device ,

[0037] FIG 8 shows a third embodiment of a beam shi fting device .

[0038] DETAILED DESCRIPTION

[0039] It might be mentioned that dimensions and proportions in the figures are typically not true to scale but exaggerated in order to clari fy procedures and ef fects .

[0040] FIG 1 shows an arrangement of an imaging system 100 for camera-based imaging of the obj ect 200 as well as the obj ect 200 to be scanned and inspected, respectively, by the imaging system 100 . The imaging system 100 is configured to generate an image IMA of the obj ect 200 , wherein the number of ef fective pixels PX of the image IMA shall be higher than the number of sensor elements SE ( i ) of the sensor chip 111 of the camera 110 .

[0041] The imaging system 100 comprises a DVS camera 110 and a control system 120 , wherein the DVS camera 110 is configured and arranged such that it captures a field of view FOV positioned in a z-direction in front of the DVS camera 110 . The obj ect 200 to be imaged is arranged at least in parts in the field of view FOV of the DVS camera 110 . Thus , a light flux field LFL from the obj ect 200 to the camera 110 falls onto an optics section 112 of the DVS camera 110 which is arranged and configured such that the light flux field LFL representing the obj ect 200 is converted into a proj ection P200 of the obj ect 200 onto a light sensitive sensor chip 111 of the camera 110 .

[0042] The light flux field LFL might be caused by illuminating the obj ect 200 with an illumination system . As outlined below, di f ferent methods of illuminating the obj ect 200 might be applied . In any case and even without a dedicated illumination system, it can be assumed that a light flux field LFL from the obj ect 200 reaches the DVS camera 110 so that the obj ect ' s 200 proj ection P200 falls onto the sensor chip 111 .

[0043] The control system 120 of the imaging system 100 comprises a control unit 122 , an evaluation unit 123 , a memory 124 , and one or more processors 125 . The control system 120 is connected to the DVS camera 110 via a wired or non-wired connection 121 . The processor 125 might be utili zed by the control unit 122 and by the evaluation unit 123 to execute a respective software to perform their respective tasks of controlling the camera 110 and evaluating imaging data DAT generated by the camera 110 . The software might be stored in the memory 124 .

[0044] The control system 120 and its control unit 122 , respectively, is configured to control the DVS camera 110 with a control signal CTRL, e . g . with regard to setting of a momentary focal length f ( t ) or any other parameter of an optics section 112 of the camera 110 and / or of a sensor chip 111 of the camera 110 . Moreover, the control system 120 is configured to receive imaging data DAT from the DVS camera 110 , both via the connection 121 . Received imaging data DAT might be stored at least temporarily in the memory 124 and / or routed to the evaluation unit 123 for further processing .

[0045] The control system 120 and its evaluation unit 123 , respectively, is configured to evaluate the imaging data DAT provided by the DVS camera 110 and potentially stored in the memory 123 to derive aspired information about the obj ect 200 and about its feature FEAT , i . e . in order to generate the image IMA.

[0046] In more general terms , the control system 120 and its evaluation unit 123 , respectively, is configured to process imaging data DAT provided by the DVS camera 110 from scanning the object 200 in order to determine pixel values PXVAL(j) of pixels PX(j) of the image IMA. For that purpose, the evaluation unit 123 applies a method METH shown in FIG 2 including a stage INSPECT for generating the imaging data DAT with the DVS camera 110 and a stage EVAL of evaluating the imaging data DAT generated in the stage INSPECT. The method METH might be implemented as a respective software executed by the processor 125 of the control system 120 as indicated above. The stage INSPECT might be controlled by the control unit 122 while the stage EVAL might be performed by the evaluation unit 123. However, the distribution of such tasks to the components of the overall system can be adapted to individual requirements of a measurement to be performed by the system.

[0047] The DVS camera 110 has a body 113 which might comprise typical electronics including a light sensitive sensor chip 111 as well as an optics section 112. Just as a regular camera lens, the optics section 112 can be a housing with an arrangement of one or more optical devices 112' like lenses, apertures etc. and imaging parameters like magnification and focal length can be set as for a regular camera lens. Thus, it can be assumed that such imaging parameters are well known during operation of the DVS camera 110. Moreover, the optics section 112 is adapted and selected such that imaging the object 200 in the field of view FOV results in a projection P200 of the object 200 or of parts of the object 200 onto the sensor chip 111. Especially, the optics section 112 might be suitable to be configured to manipulate the lateral position of the projection P200 of the object 200 relative to the sensor chip 111 as introduced in EP22199206.8 in the context of a beam shifting approach.

[0048] The sensor chip 111 comprises a plurality of light sensitive sensor elements SE(i) with i=l,...,I and I representing the total number of sensor elements SE(i) , e.g. I>106. As a simplified example, a top view of a sensor chip 111 with only 16*24=384 sensor elements SE(i) is shown in FIG 3. For the sake of clarity, only a few of the sensor elements SE(i) have been marked with reference signs. Preferably, the sensor chip 111 is a rectangular chip 111 with its sensor elements SE(i) arranged in rows and columns perpendicular to each other. The extensions of the planar, essentially two-dimensional sensor chip 111 define a coordinate system x, y, z, wherein the sensor chip's 111 plane defines the x-direction in a first dimension, e.g. along the rows, as well as the y-direction in a second dimension, e.g. along the columns, while the z- direction in the third dimension is perpendicular to the sensor chip's 111 plane and surface lllp, respectively. Individual sensor elements SE(i) have a width wPXx and a height wPXy, e.g. wPXx=wPXy=wPX . In x-direction, the distance of the centers of neighboring sensor elements SE(i) , SE(i+l) , i.e. the x-pixel-pitch, amounts to dPXx . The corresponding distance in y-direction, i.e. the y-pixel-pitch, is dPXy. Preferably, dPXx=dPXy=dPX is applicable.

[0049] In contrast to regular cameras, a signal SIG(i) generated by a DVS camera 110 for a given sensor element SE(i) at a point t(k) in time does not directly or, for example, proportionally represent the flux or intensity of light onto the respective sensor element SE(i) , but it depends on a relationship between a light flux EX onto the sensor element SE(i) at that point t(k) in time and a light flux EX onto the same sensor element SE(i) at a preceding point t(k-l) in time. More concrete, it shows whether a light flux EX at the sensor element SE(i) for a point t(k) in time differs from a light flux EX at the same sensor element SE(i) at an earlier point t(k-l) in time by more than a given threshold THRES . In case such difference DELTAFX at the location of a particular signal element SE(i) is indeed higher than the threshold THRES, it is assumed that an "event" EV(i) has occurred at the location of that signal element SE(i) , wherein "positive" and "negative" events can be distinguished. Thus, an event EV(i) detected by a sensor element SE(i) of the sensor chip 111 is the situation in which the light flux onto that sensor element SE(i) changes by more than the threshold THRES. Therein, in case the flux EX changes to a higher flux by an amount DELTAFX with DELTAFX>THRES a positive event is assumed to have happened. A change of flux to a lower value by an amount DELTAFX with DELTAFX>THRES corresponds to a negative event .

[0050] Thus, the DVS camera 110 is configured to generate for each sensor element SE(i) individually a signal SIG(i) , in the following referred to as an "event signal" SIG(i) , in case a light intensity or light flux, respectively, onto the respective sensor element SE(i) changes by more than the given threshold THRES, e.g. 15%. Therein, the event signals SIG(i) are not assigned free floating values, but each event signal SIG(i) is selected from a predefined group GRP of event signal types, which group GRP at least includes a positive signal type POS and a negative signal type NEG according to the positive and negative events introduced above, in the following occasionally referred to as "polarity" POL. Typically, the group GRP contains only two polarities, representing a positive event with POL=POS and a negative event with POL=NEG. Just for the sake of clarity, in case the intensity change DELTAFX is less than the threshold THRES, no event is assumed to have occurred and no event signal SIG would be created at all. Thus, a polarity value like POL=0 is not applicable and not required while the two polarities POL=POS, POL=NEG are sufficient for the operation of the DVS camera 110. Consequently, event signals SIG(i) are only generated in case of significant flux changes, i.e. | DELTAFX | >THRES .

[0051] More concrete, considering a specific sensor element SE(i) of the sensor chip 111 of the DVS camera 110: In case the momentary flux FX(t(k) ) onto the sensor element SE(i) is higher than the earlier flux FX(t(k-l) ) onto the same sensor element SE(i) by more than THRES, the sensor element SE(i) will generate an event signal SIG(i)=POS of positive polarity. In case the momentary flux FX(t(k) ) onto the sensor element SE(i) is lower than the earlier flux FX(t(k-l) ) onto the same sensor element SE(i) by more than THRES, the sensor element SE(i) will generate an event signal SIG(i)=NEG of negative polarity. In all other cases, i.e. as long as | DELTAFX | <THRES is applicable, no signal and no "imaging data" DAT at all are generated for the respective sensor element SE ( i ) .

[0052] Any event signals SIG(i) generated by the sensor elements SE(i) during the stage INSPECT, i.e. either positive POS or negative signals NEG, are transferred as part of the imaging data DAT(i) to the control system 120 for further processing in the stage EVAL in the evaluation unit 123. Imaging data DAT(i) transferred from the DVS camera 110 to the control system 120 for further processing include the event signal SIG(i) itself, i.e. the polarity POL(i) of the respective event EV(i) , an accurate time stamp TS (i) which stands for the point in time at which the sensor element SE(i) generated the event signal SIG(i) , and an unambiguous identifier ID(SE(i) ) of the respective individual event signal SIG(i) generating sensor element SE(i) to ensure that the position POS (SE(i) ) of such event signal SIG(i) generating sensor element SE(i) on the sensor chip 111 is known.

[0053] In the end, in contrast to a regular camera, the DVS camera does not generate frames in a regular fashion according to a frame rate, but it immediately provides the output DAT(i) to the evaluation unit 123 as soon as the respective event EV(i) occurred. Thus, the emergence of an event EV(i) is the trigger for generating and providing the respective imaging data DAT(i) , without any binding to or limitation by a frame rate. The executions of the stages INSPECT and EVAL of the method METH can be performed in parallel or at least overlapping in time because imaging data DAT(i) are transmitted immediately after they have been generated so that immediate evaluation is possible. It is not required to wait until a whole frame has been generated or all sensor elements of the chip 111 have been read out. The functional principle of the DVS camera 110 necessitates changes of the intensity or light flux FX of the object 200 projection P200 onto the sensor chip 111 and its sensor elements SE(i) . Such a change DELTAFX in flux FX onto the sensor chip 111 can be caused by different incidents, as described in PCT / EP2022 / 070785 and PCT / EP2022 / 074649. Furthermore, the beam shifting approach, e.g. according to EP22199206.8 , causes a flux change in a reproducible way, i.e. in a predefined, well known manner. The beam shifting approach achieves a relative lateral movement of the sensor chip 111 and the projection P200 of the object 200. For that, it foresees a lateral movement of the projection P200 of the object 200 onto the sensor chip 111 relative to the sensor chip 111 by a shifting amount DS. Correspondingly, the projection PFEAT of a feature FEAT of the object 200 is also shifted by the shifting amount DS since such feature projection PFEAT is simply a part of the overall object projection P200.

[0054] The relative lateral movement of the sensor chip 111 and the projection P200 can be achieved by shifting the sensor chip 111 relative to the object 200 and / or by shifting the projection P200 relative to the sensor chip 111. Shifting the sensor chip 111 relative to the object 200 can be achieved by shifting the sensor chip 111 within the camera 110 and / or by shifting the whole camera 110 with the sensor chip 111. Shifting the projection P200 relative to the sensor chip 111 can be achieved by shifting the whole object 200 and with it its original light flux field LFL and / or by controlling the optics section 112 to manipulate the projection P200. The latter is explained in the context of FIGs 5-7.

[0055] In any case, a "beam shifting device" 130 can be foreseen to be arranged and configured to perform one or more of the solutions introduced above or any other approach for achieving the relative lateral movement of the sensor chip 111 and the projection P200. For the invention introduced herein, the concrete implementation of the beam shifting device 130 is not relevant. Preferably, the sensor chip 111 is shifted, e.g. by means of a piezo drive, or the projection P220 is shifted by means of manipulating one or more components the optics section 112. In that case, the beam shifting device 130 might be integrated into the housing of the DVS camera 110, as indicated in FIG 1. However, instead of that the beam shifting device 130 might also be a separate component .

[0056] Independent from the concrete embodiment of the beam shifting device 130, the control unit 122 of the control system 120 of the imaging system 100 might be configured and connected to the respective beam shifting device 130 to control the beam shifting device 130 such that the aspired relative lateral shift is achieved in a controlled way. In other words, the shifting manner MAN can be assumed to be well known, i.e. the speed, the direction, and the shifting amount are known.

[0057] In summary, the relative movement of the projection P200 of the object 200 and the sensor chip 111 can be caused by moving the object 200, moving the projection P200, and / or moving the sensor chip 111. In any case, it is assumed that a relative lateral movement of the projection P200 and the sensor chip 111 is effected in some way, preferably controlled by the control unit 122. Therein, it can be assumed that one or possibly more shifting sequences SEQ(k) with k=l,...,K and K representing the total number of shifting sequences are executed in the stage INSPECT. In each one of the shifting sequences SEQ(k) , a relative lateral shift of the projection P200 and the sensor chip 111 by an amount DSEQ(k) and in a shifting direction DIRSEQ(k) is performed which results in a spatio-temporal shifting trajectory ST(k) . Preferably, the overall shifting amount DSEQ(k) of a shifting sequence SEQ(k) is larger than the sensor chip's pixel pitch (dPXx, dPXy) in the respective shifting direction. Shifting by one pixel pitch would in principle be sufficient. In case more than one shifting sequence SEQ(k) is foreseen, different shifting sequences SEQ(kl) , SEQ(k2) with kl^k2 differ from each other with regard to the respective shifting directions DIRSEQ ( kl ) ^DIRSEQ ( k2 ) . For example, a shifting direction DIRSEQ (k) might be along the rows or along the columns of sensor elements SE(i) of the sensor chip 111. However, in case the sensor elements SE are not arranged in a rectangular fashion, the shifting direction might be along an axis of symmetry of the sensor element arrangement.

[0058] In the following, a shifting sequence SEQ(k) is considered exemplarily which is executed such that the projection P200 moves relative to the sensor chip 111 along the rows if sensor elements, i.e. in positive x-direction. A subsequent shifting sequence SEQ(k+l) might be executed such that the projection P200 moves relative to the sensor chip 111 along the columns of sensor elements SE(i) , i.e. in positive y- direction. However, the explanations regarding shifting sequence SEQ(k) can be applied in an equivalent way and without more ado for any further shifting sequence in any direction. Therefore, only the exemplary shifting sequence SEQ(k) in positive x-direction is explained.

[0059] For the following explanations, it is assumed that the object 200 has a point-like feature FEAT, the projection PFEAT of which has an essentially Gaussian-like spatial intensity profile in the x-direction, principally comparable to the curve "FX(il,t)" shown in the upper diagram of FIG 4, i.e. in FIG 4a. Consequently, when that profile PFEAT moves across a particular sensor element SE(il) , the light flux FX(il) onto sensor element SE(il) changes over time t as shown in FIG 4a.

[0060] In response to that change of light flux FX(il,t) , the sensor element SE(il) generates an event EV(il) and respective imaging data DAT(il) , as soon as the light flux onto SE(il) has changed by more than the threshold THRES . For example, in case the original light flux FX(il,T0) at a starting point TO in time amounts to a value of 10AU (arbitrary units) and the threshold is selected to be THRES=30%, a first event would be generated at a point T1 in time, when the light flux has risen to be 13AU. The respective imaging data DAT(il) would include the time stamp TS=T1 and the polarity POL which is, as depicted in FIG 4b with an arrow pointing upwards, positive POS due to the increase of light flux. The next event would be generated at a point T2 in time, when the light flux has increased by another 30% to be 16,9AU.

[0061] As can be seen in FIG 4b, a series of positive POS events is generated when the profile PFEAT moves across a sensor element SE(il) . However, as soon as the profile's maximum has passed, the light flux EX onto SE(il) will decrease, resulting in imaging data DAT with negative polarity NEG and corresponding time stamps, each time the light flux has decreased by THRES=30%. This is again shown in FIG 4b, wherein negative polarities NEG are depicted with arrows pointing downwards.

[0062] The availability of polarities with associated time stamps and knowledge about the threshold THRES allows to reconstruct the temporal behaviour of light flux onto sensor element SE(il) to generate a light flux representation

[0063] FXrec(SE ( il ) , t , k) from the imaging data DAT(il,k) . Therein, the light flux representation FXrec(SE ( il ) , t , k) represents the temporal development of light flux FX(t) caused by the projection P200 onto the respective sensor element SE(il) during the shifting sequence SEQ(k) .

[0064] As shown in FIG 4c, the light flux representation

[0065] FXrec(SE ( il ) , t , k) for sensor element SE(i) and shifting sequence SEQ(k) is generated from the polarities POL(il,k,p) and respective time stamps TS (il,k,p) of the imaging data DAT(il,k) from an arbitrary starting value VALO, e.g. VAL0=10, by stepwise increase, preferably by a given factor FACT, preferably corresponding to the threshold THRES, e.g. FACT=THRES, in case of positive polarity POL(il,k,p) , i.e. VAL (n+1 ) =VAL (n) *FACT, and stepwise decrease, preferably by the given factor FACT, in case of negative polarity POL(il,k,p) , i.e. VAL (m+1 ) =VAL (m) / FACT . The resulting values VAL(p) are associated to the time stamps TS (p) which correspond to the polarities POL(p) which were used to determine the values VAL(p) . Due to the stepwise approach, the light flux representation FXrec(SE ( il ) , t , k) is not a continuous function, but it consists of a plurality of discrete data points, marked in FIG 4c by a cross. However, the position Textrof the maximum of the light flux representation FXrec(SE ( il ) , t , k) in time is an indication for the point in time at which the originating projection PFEAT was positioned right on the sensor element SE(il) . Thus, from the knowledge of Textrand considering the known shifting manner MAN, the starting position of that projection PFEAT before the shifting sequence SEQ(k) has started can be reproduced. That starting position can be associated with a pixel PX(jl) position in the high resolution image IMA.

[0066] This high resolution image IMA can be congruent with the sensor chip 111, but is has more intermediate effective pixels PX(j) both in x- and y-direction. This means that the image IMA builds, on the one hand, on the fixed pattern of sensor elements SE of the chip 111. On the other hand, new sampling points and pixels PX, respectively, are created due to the occurrence of events EV, depending on the light flux distribution of the projection P200 during shifting. Thus, a static way of pixel interpolation across the sensor 111, e.g. with a constant factor of 4, 6, or 9 etc., is not necessarily applicable. However, the accuracy of determination of Textrallows to determine the starting position with an accuracy which is better than the pixel pitch of the sensor chip 111. This allows to generate the image IMA with a number of effective pixels PX which is higher than the number of sensor elements SE(i) of the sensor chip 111.

[0067] After having determined the identity of the pixel PX(jl) of the image IMA, the pixel value PXVAL(jl) to be assigned to that pixel PX(jl) can determined to be the value VAL of the maximum of the light flux representation FXrec(SE ( il ) , t , k) , i.e. PXVAL(jl) = FXrec(SE (il) , Textr, k) .

[0068] However, the nature of the light flux representation FXrec(SE ( il ) , t , k) consisting of a plurality of discrete data points allows only a limited accuracy of determination of Textr. Accuracy can be improved by further processing the light flux representation FXrec(SE ( il ) , t , k) by fitting or interpolating, respectively, the light flux representation FXrec(SE ( il ) , t , k) to generate an interpolated light flux representation FXrec, fit (SE ( il ) , t , k) , as shown in FIG 4c with the dotted line curve. In contrast to FXrec(SE ( il ) , t , k) , the interpolated curve FXrec,fit(SE ( il ) , t , k) is a continuous function so that the position Textrof the maximum of the interpolated light flux representation FXrec, fit (SE ( il ) , t , k) in time can be determined with higher accuracy. As before, Textris an indication for the point in time at which the originating projection PFEAT was positioned right on the sensor element SE(il) . Thus, from the knowledge of Textrand considering the known shifting manner MAN, the starting position of that projection PFEAT before the shifting sequence SEQ(k) has started can be reproduced. That starting position can be associated with a pixel PX(jl) position in the high resolution image IMA. The corresponding pixel value PXVAL(jl) of that pixel PX(jl) can again be determined to be the maximum value of the interpolated curve, i.e.

[0069] PXVAL ( j 1 ) =EXreC; fit( SE ( i 1 ) ,Textr,k) .

[0070] Thus, execution of the above approach, at least for selected signal elements SE(il) for which a plurality of events has been generated during the shifting sequence SEQ(k) , allows to identify pixels PX(jl) and their values PXVAL(jl) of the high resolution image IMA. In principle, all signal elements SE(i) of the sensor chip 111 can be selected signal elements SE(il) . However, preferably only those ones of the signal elements SE(i) are selected signal elements SE(il) , the imaging data DAT(il) of which correspond to a predefined pattern. Such a predefined pattern might be characterized by a series of at last PLU>1 polarities POL(i,k) of the same kind, i.e. either POS or NEG, in a row. The example of FIG 4 include a series of eight positive polarities, followed by s series of six negative polarities. In principle, PLU=2 would be sufficient to apply the approach introduced herein, but higher PLU would be beneficial.

[0071] In case the feature FEAT is not only a point, i.e. a structure for which the light flux increases and decreases again or vice versa, but, for example a transition from a continuous high-flux region to a continuous low-flux region, i.e. an edge, the light flux develops as shown in FIG 5. FIG 5 shows an exemplary light flux FX(t) onto signal element SE(il) over time during the shifting sequence SEQ(k) . The projection PFEAT is assumed to correspond to an edge like feature FEAT extending at the object 200 along the y- direction. Although the edge FEAT itself might be a sharp structure, the corresponding projection PFEAT will not include a discrete transition from high to low light flux EX, but it will be characterized by a certain blurriness and a continuous transition with a certain slope. The form of the transition is characteristic for the feature FEAT and, at the same time, depends on the optical imaging properties of the imaging system 100, i.e. of the DVS camera 110, the optics section 112, the beam shifting device 130 and other components of the imaging system 100 which influence the procedure of projecting the object 200 onto sensor chip 111. However, those optical properties can be assumed to be well known so that it is possible to derive the real physical form of a feature FEAT, i.e. its dimensions and extension, from the projection PFEAT.

[0072] However, in this scenario the approach described above would not be applied for the light flux representation FXrec, fit (SE ( il ) , t , k) itself, but for its derivative dFXrec, fit (SE ( il ) , t , k) / dt . I.e. the light flux representation FXrec(SE ( il ) , t , k) and the interpolation FXrec,fit(SE ( il ) , t , k) are determined as described above, but the point Textr in time is the position of the maximum of the time derivative dFXrec, fit (SE ( il ) , t , k) / dt . The pixel position PX(jl) is determined based on Textr and MAN as described above. The respective pixel value PXVAL(jl) might be, for example, the value of the derivative for t=Textr.

[0073] Thus, the DVS camera 110 with the beam shifting device 130 can be used for precise dimensional measurements, when the timing information of the beam shifting device 130 and the DVS camera 110 gets exploited. When the trajectory of the beam shifting, i.e. the path of the projection P200 on the sensor chip 111 expressed by the "shifting manner" MAN, is known, each imaging data DAT(i) can be attributed to a specific phase of the beam shifting pattern. Thus, the effect of the shifting can be compensated for by evaluating the temporal and spatial event signal pattern expressed by the event polarities, i.e. "POS" or "NEG", and the event threshold THRES . By evaluating the event steps in the time series of event signals SIG(il) from one sensor element SE(il) or, possibly, a group of sensor elements SE covered by the beam shifting, and subsequent interpolation, a spatial- temporal pattern FXrec,fit can be deduced. The position of intensity maxima, minima, or maximum slope for a rising or falling signal of FXrec,fit can be determined from this spatial temporal pattern and assigned to one reference point TO of time for the shifting path, where no deflection has happened yet. When all the lateral positions of events EV(il) , EV(i2) , EV(i3) , etc. get assigned with respect to TO, they get referenced to the same point in time and, thus, pixel pattern position. Due to the timing information and higher temporal resolution due to the interpolation, the feature points like maxima, minima, or points of maximal slope can be assigned based on Textr with sub-pixel accuracy and thus the beam shifting technology for the DVS camera 110 allows to increase the physical resolution of the DVS camera sensor 111 based on the timing information for the event signals and the known trajectory of the shifting, expressed by MAN. Thus, the plurality of event signals SIG(il) and POL(il) , respectively, generated during the shifting sequence or time series resulting from the relatively blurry projection PFEAT being shifted across a particular sensor element SE(il) as explained above are assigned to the corresponding time stamps TS (il) . The resulting temporal behavior FXrec(il,t) , derived from those discrete data points, is a profile PROF of the projection PFEAT made up by a plurality of discrete data points from which the real physical form of a feature FEAT can be reproduced based on the knowledge about the optical imaging properties of the imaging system 100 and the knowledge of the shifting manner MAN. Sub-pixel resolution of the real physical form of a feature FEAT is achieved by using the discrete data points as input data for an interpolation of the projection PFEAT.

[0074] The beam shifting device 130 can be embodied in different ways .

[0075] For example and as shown in FIG 5, the beam shifting device 130 can be embodied as a transparent, piano-parallel plate, which has two parallel surfaces 131sl, 131s2 and which is positioned in a z-direction between the object 200 and the sensor chip 111, so that the light flux field LFL emitted by the object 200 passes through the parallel surfaces 131sl, 131s2 and through the inner of the plate 131 to reach the sensor chip 111, wherein the lateral shifting is achieved by tilting the plate 130 around the x-direction and / or around the y-direction.

[0076] In another example, as shown in FIG 6, the beam shifting device can be embodied as a pair of mirrors 132 which are arranged parallel to each other and, in a default position, under an angle of essentially 45° with regard to the z- direction, such that the light flux field LFL emitted by the object 200 is reflected by a first one of the mirrors 132 to the second one of the mirrors 132 and from the second one of the mirrors 132 towards the sensor chip 111, wherein the lateral shi fting is achieved by tilting both mirrors by equal angels ALPHA around the x-direction and / or around the y- direction relative to the default position .

[0077] In a further example , shown in FIG 7 , the beam shi fting device 130 can be embodied as an electro-optical device which has a plate 133 with two surfaces 134 and which is positioned in a z-direction between the obj ect 200 and the sensor chip 111 , so that the light flux field LFL emitted by the obj ect 200 passes through the surfaces 134 and through the inner of the plate 133 to reach the sensor chip 111 , wherein the lateral shi fting is achieved by applying a voltage at preferably transparent electrodes arranged at the surfaces 134 .

[0078] For each one of those embodiments of the beam shi fting device 130 , the respective elements might be integrated into the optics section 112 . In any case , the beam shi fting device 130 would preferably be controlled by the control unit 122 .

[0079] Preferably, the imaging system 100 includes an illumination system with one or more light sources 131 , e . g . LEDs , wherein the illumination system and its one or more light sources 131 are configured and arranged such that they provide a preferably constant light flux onto the obj ect 200 which is preferably homogenous over the entire field of view FOV . Since typical DVS cameras 110 are monochromatic, the illumination can be monochromatic too or it is a multicolor illumination, e . g . in the ultraviolet , short wavelength infrared, or in the visible spectrum . In general , the illumination is adapted to the spectral sensitivity range of the respective DVS camera and can therefore range from ultraviolet to visible to infrared including SWIR, NIR spectral ranges , i . e . from < 190 nm up to 2 , 5 pm . All known and typically applied illumination modalities can be used individually or in a combined manner, like bright field transmissive light illumination, bright field reflected light , dark field reflected light or dark field transmitted light , as well as any kind of mixed forms or modalities , e . g . a combination of transmissive and reflective light . Further contrasting methods like polari zation might also be combined for the illumination . In any case , natural and / or arti ficial illumination of the obj ect 200 achieves that the light flux field LFL is emitted by the obj ect 200 towards the DVS camera 110 which can then detect the light flux field LFL . Therein, ambient light might be suf ficient to cause the light flux field LFL, but utili zation of dedicated light sources 131 and an illumination system, respectively, is preferred because this allows generation of well defined, homogeneous illumination .

[0080] Since DVS cameras detect changes in the flux FX per sensor element SE and generate an event i f the photocurrent rises or lowers by a pre-defined amount THRES or more , they automatically adapt themselves to the actual irradiance level for each sensor element SE in the operating range of the DVS camera 110 with respect to speed given by the minimal read out time of the sensor elements ' SE photocurrent and the maximum time for this evaluation . In concrete , but to be understood as an example , this might result in an ef fective dynamic range of more than l O OdB or even 120dB and above , compared to about 40 to 50dB for standard CMOS or CCD cameras with full frame read out . This capability is a result of the DVS sensor chip' s 111 intrinsic ability to adapt locally to the respective signal strengths so that it applies the full well capacity of a sensor element SE with di f ferent exposure times where the respective sensor element is monitoring for a change in light exposure . In ef fect this makes up the ef fective dynamic range of l O OdB and above .

[0081] In principle , the usage of events instead of a broad range of values representing fluxes or similar acts as a first kind of data or event filter to reduce the amount of measured data right at the source , i . e . at the imaging system 100 itsel f . In a very broad and general perspective , the approach based on usage of a DVS camera 110 helps to directly reduce the data created to the relevant portion without long chains of image processing and thus also to improve the imaging system 100 performance with respect to sustainability aspects like energy consumption, materials involved due to less powerful computing and data transmission systems , and overall ICT infrastructure and also related carbon footprint . Despite these optimi zations , the core of the imaging in finding potential defect sites is much more ef ficient and detailed than by employing conventional cameras .

[0082] From a data point of view and independent from the concrete embodiment of the obj ect 200 and independent from the scenario , the amount of data created from the imaging system 100 with a DVS camera 110 is decreased signi ficantly as compared to a conventional camera, since all regions of the obj ect 200 without relevant features FEAT do not cause imaging data DAT ( i ) because the changes in light flux of the respective illumination modalities do not exceed the predefined threshold THRES .

[0083] Thus , the solution allows to utili ze the advantages of the DVS camera 110 , e . g . speed, high dynamic range , and data volume reduced to the minimum, to generate an image IMA of the obj ect 200 which depicts the features FEAT of the obj ect 200 , being edges or contour lines etc . , i . e . intrinsic obj ect features , and / or irregularities . The invention can be performed with obj ects at rest or moving relative to the camera 110 .

[0084] While the present invention has been described above by reference to various embodiments , it should be understood that many changes and modi fications can be made to the described embodiments . It is therefore intended that the foregoing description be regarded as illustrative rather than limiting, and that it be understood that all equivalents and / or combinations of embodiments are intended to be included in this description . Thus , the invention is not restricted to the above illustrated embodiments , but variations can be derived by a person skilled in the art without deviation from the scope of the invention .

Claims

Claims1. Method METH for camera based imaging of an object (200) for generating a high-resolution image IMA of the object (200) with an imaging system (100) , wherein the imaging system (100) comprises-the camera (110) , which is embodied as a DVS camera (110) with a light sensitive sensor chip (111) with a plurality of sensor elements SE(i) , each of which generates imaging data DAT(i) in case of an event EV(i) of change of light flux EX onto that sensor element SE(i) by more than a given threshold THRES,- a beam shifting device for effecting a relative lateral shift of the sensor chip (111) and a projection P200 of at least a section of the object (200) onto the sensor chip (111) , wherein the method METH comprises-a stage INSPECT of observing the object (200) with the DVS camera (110) , including execution of one or more shifting sequences SEQ(k) , wherein during each shifting sequence SEQ(k) a relative lateral shifting of the projection P200 and the sensor chip (111) is executed by means of the beam shifting device (130) in a known shifting manner MAN to generate imaging data DAT(i,k) ,-a stage EVAL of evaluating the imaging data DAT(i,k) for selected ones SE(il) of the sensor elements SE(i) , wherein o for each selected sensor element SE(il) , a light flux representation FXrec( SE ( il ) , t , k) is generated from the imaging data DAT(il,k) , and o pixel values PXVAL(j) of pixels PX(j) of the image IMA are determined based on the light flux representations FXrec(SE (il) , t, k) .

2. Method according to claim 1, wherein the light flux representation FXrec(SE ( il ) , t , k) is generated from the polarities POL(il,k,p) and respective time stamps TS (il,k,p)included in the imaging data DAT(il,k) from a starting value VALO by stepwise- increase, preferably by a given factor FACT, preferably corresponding to the threshold THRES, in case of positive polarity POL(il,k,p) and- decrease, preferably by the given factor FACT, in case of negative polarity POL(il,k,p) .

3. Method according to any one of claims 1 to 2, wherein the light flux representation FXrec(SE ( il ) , t , k) is further processed to determine- the point Textr in time at which the light flux representation FXrec( SE ( il ) , t , k) or its derivative EX'rec( SE ( il ) , t , k) in time has an extreme value and- the respective value EXT=FXrec(SE ( il ) , Textr, k) itself.

4. Method according to claim 3, wherein the identity of the pixel PX(jl) , for which a pixel value PXVAL(jl) is to be determined, is derived from the point Textrin time and from the known shifting manner MAN.

5. Method according to any one of claims 3 to 4, wherein the value EXT is used to determine the pixel value PXVAL(jl) of the pixel PX(jl) , preferably such that PXVAL(jl) is proportional to EXT, especially PXVAL ( j 1 ) =EXT .

6. Method according to any one of claims 3 to 5, wherein the further processing of the light flux representationFXrec(SE ( il ) , t , k) includes fitting the light flux representation FXrec(SE ( il ) , t , k) to determine an interpolated light flux representation FXrec, fit (SE ( il ) , t , k) , wherein- Textris the point in time at which the interpolated light flux representation EXrec, fit ( SE ( il ) , t , k) or its derivative EX' rec, fit (SE (il) , t, k) = dFXrec, fit ( SE ( i 1 ) , t , k) / dt has an extreme value and- the respective value EXT is the value FXrec, fit ( SE (ID , Textr, k) Itself.

7. Method according to any one of claims 1 to 6, wherein all signal elements SE(i) are selected signal elements SE(il) .

8. Method according to any one of claims 1 to 6, wherein only those ones SE(il) of the signal elements SE(i) are selected signal elements SE(il) , the imaging data DAT(il) of which correspond to a predefined pattern.

9. Method according to claim 8, wherein the predefined pattern is characterized by a series of at last PLU>1 polarities POL(i,k) of the same kind in a row.

10. Method according to any one of claims 1 to 9, wherein the stage INSPECT includes a plurality of shifting sequences SEQ(k) with k=l,...,K and K>1, wherein different shifting sequences SEQ(kl) , SEQ(k2) with klVk2 differ from each other with regard to the respective shifting directionsDIRSEQ (kl) VDIRSEQ (k2) .

11. Method according to any one of claims 1 to 10, wherein the lateral shifting during each shifting sequence SEQ(k) is executed such that the projection P200 moves relative to the sensor chip (111) along a straight line, preferably along columns or rows of the sensor elements SE(i) on the sensor chip (111) .

12. Method according to any one of claims 1 to 11, wherein the relative lateral shifting of the projection P200 and the sensor chip (111) is executed with the help of the beam shifting device (130) by- shifting the projection P200 relative to the sensor chip (111) , preferably by controlling an optics section of the DVS camera, and / or- shifting the sensor chip (111) relative to the projection P200.

13. Method according to claim 12, wherein the optics sectionoptical devices ( 112 ' ) for generating the projection P200 of the object (200) from the light flux field LFL emitted by the object (200) , wherein the beam shifting device (150) is realized by at least one of the optical devices ( 112 ’ ) of the optics section (112) , wherein the lateral shifting is achieved by changing optical properties of the at least one optical device (112' ) , preferably by shifting, tilting, and / or deforming the at least one optical device ( 112 ’ ) .

14. Imaging system (100) for camera (110) based imaging of an object (200) for generating a high-resolution image IMA of the object (200) , configured to execute a method METH according to any one of claims 1 to 13 and comprising-the camera (110) , which is embodied as a DVS camera (110) with a light sensitive sensor chip (111) with a plurality of sensor elements SE(i) , each of which generates imaging data DAT(i) in case of an event EV(i) of change of light flux EX onto that sensor element SE(i) by more than a given threshold THRES,-the beam shifting device (130) for effecting a relative lateral shift of the sensor chip (111) and a projection P200 of at least a section of the object (200) onto the sensor chip (111) , and-a control system (120) with o a control unit (122) for controlling the beam shifting device and the sensor chip to generate the imaging data DAT ( i ) and o an evaluation unit (123) for executing the stage EVAL for evaluating the imaging data DAT(i) to generate the high- resolution image IMA.

15. Imaging system (100) according to claim 14, wherein the DVS camera (110) comprises an optics section (112) for converting a light flux field LFL from the object (200) into the projection P200 of the object (200) onto the sensor chip (111) , wherein the optics section (112) comprises one or more optical devices ( 112 ’ ) for generating the projection P200 from the light flux field LFL, wherein-the beam shifting device (130) is realized by at least one of the optical devices ( 112 ’ ) of the optics section (112) ,- the lateral shifting is achieved by changing electrical and / or optical properties of the at least one optical device ( 112 ’ ) , preferably by shifting, tilting, and / or deforming the at least one optical device (112' ) .

Images