Image Sensor

Abstract

To provide an imaging system having functionality for detecting faults in the control and / or readout of an image sensor.SOLUTION: An image sensor included in an imaging system 300 comprises a plurality of imaging pixels and at least one diagnostic pixel (represented by shaded boxes) coupled to a reference signal. The reference signal of the diagnostic pixel is generated using a potential divider network RA-RC connected to a reference voltage Vref. The imaging system comprises control and readout circuitry 320 configured to: select one or more of the imaging pixels and, for each selected imaging pixel, read out an imaging signal; select the at least one diagnostic pixel and, for each selected diagnostic pixel, read out a diagnostic signal that is indicative of the reference signal of the selected diagnostic pixel; and output the diagnostic signal for use in system diagnostics.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present disclosure relates to an image sensor system comprising an image sensor having one or more diagnostic pixels and control and readout circuitry arranged to read out diagnostic signals from the diagnostic pixels. [Background technology]

[0002] An image sensor typically comprises a plurality of imaging pixels arranged in a matrix of rows and columns. Each pixel accumulates a charge depending on the light incident on the pixel. Typically, the charge is output from the image sensor using readout lines and then converted to digital using an analog-to-digital converter (ADC). For example, a control circuit may select pixels to be readout using control lines that span the image sensor in one dimension (e.g., there may be one or more control lines running along each row so that pixels can be selected row by row). The charges on the selected pixels may then be output using readout lines that span the image sensor in a second dimension (e.g., there may be at least one readout line running along each column so that pixels in a selected row can be output simultaneously using the readout lines). Those output values ​​may then be converted to digital by one or more ADCs.

[0003] Image sensors are used for a variety of different purposes, many of which require the accuracy and reliability of the image sensor for safety reasons. For example, autonomous or semi-autonomous vehicles may use time-of-flight (ToF) or lidar (LIDAR) systems to understand the environment around the vehicle and make decisions about how the vehicle can safely proceed. ToF and LIDAR systems typically have an image sensor used to generate a depth map / frame of the scene, which the vehicle can use to understand the 3D positions of objects around the vehicle. If the image sensor is not functioning properly, the depth map / frame may contain errors or inaccuracies that can result in the vehicle taking unsafe actions. In a further example, 2D image sensors (i.e., image sensors that are not ToF or LIDAR) may be used for safety-critical purposes, such as for object recognition in safety-critical systems, where an operational error in the image sensor could cause unsafe actions.

[0004] There are many different types of faults that can occur with an image sensor and its surrounding control and readout circuitry. For example, rows may be selected and read out in the wrong order, which can result in an image that inaccurately represents the imaged scene, as some rows of the image may appear to be imaging one part of the scene when in fact they are imaging another part of the scene. Another error is that columns may be read out of the wrong order, which can result in an image that inaccurately represents the imaged scene, as some columns of the image may appear to be imaging one part of the scene when in fact they are imaging another part of the scene. Another error can be a fault in the row select control line and / or charge readout line, such as a line break, which can result in an image that includes pixels that are not properly selected and / or readout, such that the content of these pixels is unreliable. Another error can be a fault in one or more ADCs, resulting in inaccurate and therefore unreliable pixels in the image. When an imaging sensor is used in a safety-critical system, it may be important to detect when such a fault is occurring so that the system using the output of the imaging sensor can take appropriate action, such as stopping use of the imaging sensor, or powering off the safety-critical system, or communicating that a fault has occurred so that it can be investigated and repaired. Summary of the Invention [Means for solving the problem]

[0005] One aspect of the present disclosure provides an imaging system, comprising: 1. An image sensor, comprising: a plurality of imaging pixels arranged in a matrix of rows and columns, each imaging pixel configured to accumulate charge dependent on light incident on the imaging pixel; an image sensor comprising: at least one diagnostic pixel coupled to a reference signal; A control and readout circuit, selecting one or more of the imaging pixels and, for each selected imaging pixel, reading an imaging signal indicative of the charge accumulated on the selected imaging pixel; selecting at least one diagnostic pixel and reading, for each selected diagnostic pixel, a diagnostic signal indicative of the reference signal of the selected diagnostic pixel; and a control and readout circuit configured to output a diagnostic signal for use in system diagnostics.

[0006] Another aspect of the present disclosure provides an imaging system, comprising: 1. An image sensor, comprising: a plurality of imaging pixels arranged in a matrix of rows and columns, each imaging pixel configured to accumulate charge dependent on light incident on the imaging pixel; an image sensor comprising, for each row of imaging pixels, a row control line for selecting the row of imaging pixels for readout; a row control circuit coupled to each row control line and configured to output a control signal to each row control line for controlling the selection and readout of the imaging pixels; and diagnostic circuitry coupled to each row control line and configured to output a signal generated based on the control signal on the row control line during selection and readout of the imaging pixels.

[0007] Another aspect of the present disclosure provides an imaging system, comprising: 1. An image sensor, comprising: a plurality of imaging pixels arranged in a matrix of rows and columns, each imaging pixel configured to accumulate charge dependent on light incident on the imaging pixel; an image sensor comprising: a first plurality of diagnostic imaging pixels, some of which are shaded such that they do not accumulate charge dependent on light incident on the diagnostic imaging pixel, and other of which are not shaded such that they accumulate charge dependent on light incident on the diagnostic imaging pixel; A control and readout circuit, selecting one or more of the imaging pixels and, for each selected imaging pixel, reading an imaging signal indicative of the charge accumulated on the selected imaging pixel; selecting at least one of the plurality of diagnostic imaging pixels and, for each selected diagnostic pixel, reading out a diagnostic signal indicative of the charge accumulated on the selected diagnostic pixel; and a control and readout circuit configured to output a diagnostic signal for use in system diagnostics. [Brief explanation of the drawings]

[0008] Aspects of the present disclosure will now be described, by way of example only, with reference to the following drawings, in which:

[0009] [Figure 1] 1 illustrates the structure of an exemplary imaging pixel. [Figure 2] 1 illustrates the structure of an exemplary diagnostic pixel. [Figure 3] 1 shows an exemplary schematic diagram of an imaging system according to an aspect of the present disclosure. [Figure 4] 4 shows an exemplary schematic diagram of further details of the imaging system of FIG. 3. [Figure 5] 1 shows an exemplary schematic diagram of an imaging system according to a further aspect of the present disclosure. [Figure 6] 1 shows an exemplary schematic diagram of an imaging system according to a further aspect of the present disclosure. [Figure 7] 1 shows an exemplary schematic diagram of an imaging system according to a further aspect of the present disclosure. [Figure 8] 1 shows an exemplary schematic diagram of an imaging system according to a further aspect of the present disclosure. [Figure 9] 1 shows an exemplary schematic diagram of an imaging system according to a further aspect of the present disclosure. [Figure 10] 1 shows an exemplary schematic diagram of an imaging system according to a further aspect of the present disclosure. [Figure 11] 1 shows an exemplary schematic diagram of an imaging system according to a further aspect of the present disclosure. [Figure 12]12 illustrates further details of an exemplary implementation of the system of FIG. 11. [Figure 13] 1 shows an exemplary schematic diagram of an imaging system according to a further aspect of the present disclosure. [Figure 14] 12 illustrates further details of an exemplary implementation of the system of FIG. 11. DETAILED DESCRIPTION OF THE INVENTION

[0010] When an image sensor is used as part of a safety-critical system, a common requirement is that the order of image readout is known so that it can be detected if the row and / or column readout order is incorrect (e.g., if the image is inverted, if rows are swapped, if rows are skipped, if rows are repeated, if columns are swapped, if columns are skipped, if columns are repeated, etc.). For image sensors that are not themselves certified to be functionally safe, the readout order may need to be verifiable by another device in the system to create a safety system. This means that diagnostic information used to verify the readout order should be embedded in the image data read out from the image sensor.

[0011] Some systems may attempt to digitally replace the values ​​in some image sensor pixels with predetermined diagnostic values, so that it can be detected whether those diagnostic values ​​are being read out accurately from the same locations they were set in. However, such solutions assume that the circuitry and control signals that apply the diagnostic values ​​to the image sensor are operating correctly and accurately, which is not always the case.

[0012] Accordingly, the inventors have developed a technique by which diagnostic data can be read out from an image sensor without relying on additional diagnostic circuitry (such as circuitry configured to apply diagnostic values ​​to specific pixels of the image sensor) and in a manner that allows for detection of faults in the control and / or readout circuitry of the image sensor.

[0013] Figure 1 shows an example of a typical imaging pixel structure. The pixel includes a photodiode (in this example, a pinned photodiode, but any suitable type of photodiode can be used) in which charge accumulates depending on the light incident on the pixel. Charge can be read out of the pixel onto a readout line using the control signals Reset, Select, and TX. Readout can be performed by performing the following process: 1. Enable the select signal to turn on the FET select. 2. Pulse the reset signal (e.g., turn on and then off) so that a reset voltage is placed on the floating diffusion (FD) of the pixel and output on the readout line. 3. Sample the output reset voltage, for example by converting the reset voltage to digital using an ADC coupled to the readout line and storing that digital value. 4. Pulse the signal on TX (e.g., turn it on and then off) so that the photodiode voltage is placed on the pixel's floating diffusion (FD) and output on the readout line. 5. Sample the output photodiode voltage, for example by converting the photodiode voltage to digital using an ADC coupled to the readout line and storing that digital value. 6. Determine the final pixel value using the sampled reset voltage and the sampled photodiode voltage, for example by performing correlated double sampling (CDS).

[0014] As will be appreciated by those skilled in the art, various other pixel structures and processed controls are possible, such as a differential pixel including two photodiodes, or a quad pixel including four photodiodes, and the above structure and process are merely one non-limiting example.

[0015] 2 shows an example of a diagnostic pixel according to one embodiment of the present disclosure. It can be seen that its structure is exactly the same as that of FIG. 1, except that it does not include a photodiode, but instead includes a coupling to a predetermined reference voltage. Readout can be performed in the same manner as described above. As a result, the same structure and process are used for the diagnostic pixel as for a photodiode pixel, except that if all of the control and readout circuitry is operating correctly and reliably, the final pixel value reached at the end of the readout process should correspond to the predetermined reference voltage (within an error range that accounts for factors such as signal noise and reset voltage variation).

[0016] FIG. 3 shows an exemplary schematic diagram of an imaging system 300 according to one embodiment of the present disclosure. System 300 includes an image sensor 310 with a plurality of imaging pixels (represented by unshaded boxes) arranged in a matrix of N rows and M columns. Each imaging pixel is configured to accumulate charge depending on light incident on the imaging pixel, and each pixel may be configured, for example, as shown in FIG. 1 (or in any other known manner, such as differential pixels). Image sensor 310 also includes a plurality of diagnostic pixels (represented by shaded boxes). Each diagnostic pixel may be configured in the same manner as each of the imaging pixels, except that rather than being configured to accumulate charge, each diagnostic pixel is coupled to a reference voltage, as described above in connection with FIG. 2. In this example, the plurality of diagnostic pixels are arranged as a column, column_D.

[0017] The imaging system 300 also includes a control and readout circuit 320 configured to select one or more imaging pixels and, for each selected imaging pixel, read out an imaging signal indicative of the charge accumulated on the selected imaging pixel. It is also configured to select at least one diagnostic pixel and read out a diagnostic signal indicative of a reference voltage of the selected diagnostic pixel. For these purposes, the control and readout circuit 320 may include a row control circuit 330 configured to control pixel selection and readout row by row. For this purpose, the imaging system 300 includes N row control lines 335_N, each of which may actually include multiple lines for carrying multiple control signals. For example, in the arrangement of FIG. 1, three control signals are required to control pixel selection and readout, i.e., TX, RESET, and SELECT. Thus, although each of the N row control lines 335_N is shown in FIG. 3 as a single line for clarity, in reality, they may include multiple lines, e.g., one each for TX, RESET, and SELECT. As can be seen, each row control line 335_N is shared by all of the imaging and diagnostic pixels in that row. As a result, row control unit 330 is configured to simultaneously control the selection and readout of all imaging and diagnostic pixels in a particular row. For each selected imaging pixel, an imaging signal is read out via their common column readout line 342_M. For example, imaging signals for selected imaging pixels in column_1 are read out via readout line 342_1, imaging signals for selected imaging pixels in column_2 are read out via readout line 342_2, imaging signals for selected imaging pixels in column_M are read out via readout line 342_M, and so on. Each readout imaging signal, e.g., each imaging voltage, indicates the charge accumulated on the respective imaging pixel, as described above with reference to FIG. 1.

[0018] 3, each of row control lines 335_N is also coupled to a respective one of the N diagnostic pixels, such that row control unit 330 controls the selection and readout of the diagnostic pixels in exactly the same manner as the imaging pixels. For each selected diagnostic pixel, an imaging signal is read out via lines 342_1 through 342_M, while a diagnostic signal is read out via a common column readout line 342_D. The diagnostic signal indicates a reference voltage for the selected diagnostic pixel, but may not be exactly the same as the reference voltage as a result of noise, etc.

[0019] Analog signals read out via common column readout lines may be converted to digital by analog-to-digital converters (ADCs) 340. In this example, there is one ADC for each readout line so that conversion of each of the readout signals may occur simultaneously. However, alternatively, there may be fewer ADCs than readout lines (potentially only one ADC), with the control and readout circuitry 320 also including one or more multiplexers to multiplex two or more signals to each ADC. In a further alternative, there may be more ADCs than readout lines, for example, two or more ADCs for each readout line. Digital representations 348 of the imaging and diagnostic signals are output from the control and readout circuitry 320. Alternatively, the ADC 340 may be omitted, and the readout imaging and diagnostic signals may be output from the control and readout circuitry 320 as analog signals for digital conversion elsewhere or for further processing / analysis in the analog domain.

[0020] In this example, the reference voltage for each of the diagnostic pixels is different. In the example of Figure 3, this is done by taking a reference voltage Vref (e.g., from any suitable voltage reference generator such as a bandgap voltage reference) and dividing it into potential divider networks R A ~R C This is achieved by generating N different references 355_N using a reference voltage . However, it may alternatively be achieved in any other suitable way, for example using N independent reference voltages, etc.

[0021] By setting the voltage reference differently for each diagnostic pixel, the diagnostic signal read out from each diagnostic pixel can be used as a unique row ID. For example, the control and readout circuitry 320 can read out the image sensor 310 row by row. As signals are read out from the pixels of one row, it can continue to move forward and select and read out signals from the next row until all the pixels of the row have been read out. Each row of imaging signals will be accompanied by a diagnostic signal that uniquely identifies that row. As a result, if the row control unit 330 has a fault and selects in the wrong order, for example, row 1, row 3, row 2, rather than processing in numerical order, this can be detected using the diagnostic signal read out for each row, because the rows identified by the diagnostic signal are in the wrong order. In contrast, without the diagnostic pixel of the present disclosure, the pixel data being read out from the image sensor would simply be assumed to be read out in the correct order.

[0022] To use the diagnostic signal, when the imaging signal for one row is output as a digital value 348_1-348_M, the accompanying digital value 348_D can be compared to known and predetermined voltage references 355_1-355_N. If the diagnostic signal is identical to one of the reference voltages within an acceptable range of accuracy (e.g., within a range of + / - 1% or + / - 2% of the reference voltages 355_1-355_N), then the imaging signal 348_1-348_M is known to come from the row corresponding to that reference voltage. This takes into account inherent imprecision in the diagnostic signal, for example, as a result of signal noise, imprecision in the digital conversion, and variability in the reset levels of the diagnostic pixels. The levels of the voltage references 355_1-355_N can be stored in memory, along with the corresponding row numbers, for example, so that the known and predetermined voltage reference levels and their corresponding row numbers can be retrieved for purposes of this diagnostic process. The diagnostic process may be performed by any suitable processor, such as a diagnostic unit that is part of the imaging system (or external to the imaging system), or by any other type of image processor configured to perform this diagnostic function in addition to its normal image processing functions.

[0023] Furthermore, because the diagnostic pixels use the same pixel select and readout control signals as the imaging pixels, other types of errors / faults in the imaging system can be identified using the diagnostic signals, such as determining that the select signal is stuck low for a particular row (e.g., because that would result in a diagnostic signal indicating an incorrect row ID), and / or that the select signal is stuck high for a particular row (e.g., because the diagnostic signals for all other rows are corrupted, thus indicating an incorrect row ID), and / or that there is a fault in any one of the reset and / or tx signals, and / or that the order of control signal operation is incorrect (e.g., the reset and tx signals are pulsed in the wrong order).

[0024] Further, in this example, it can be seen that the diagnostic pixel column, COLUMN_D, is located on the opposite side of image sensor 310 from where control and readout lines 335_1-335_N receive control and readout signals from row control unit 330. In particular, control and readout lines 335_1-335_N extend in rows across image sensor 310, with signals propagating from one side of image sensor 310 to the other. The diagnostic pixel column, COLUMN_D, is located on the far side of image sensor 310 from where those signals originate, which means that if there is any fault anywhere along any of control and readout lines 335_1-335_N, it will affect the diagnostic pixel and therefore will be reflected in the readout diagnostic signal.

[0025] FIG. 4 shows some further example details of the control and readout circuitry 330, specifically the circuitry related to the timing of the select signals to each row (for clarity, further circuitry related to the reset and TX signals is not shown). These details are presented merely to aid in understanding some example operational implementations of the control and readout circuitry 320 and to provide an example of how the circuitry may be configured to serially select one row at a time. In this example, a shift register is comprised of a series of cascaded flip-flops 410_1 through 410_N to shift select pulses from one row to the next using a clock signal. Those skilled in the art will readily appreciate that the timing and operation of the control signals may be achieved in many different ways, and that this is merely one particular illustrative example.

[0026] The inventors recognize that the implementation of the diagnostic pixel of Figures 3 and 4 may be difficult to realize for large imaging arrays, e.g., having 1000 or more rows. In particular, a unique reference signal would be required for each row. This means that the difference between adjacent reference signals may be very small, thereby imposing stringent requirements on accuracy levels, requiring, for example, a high-precision voltage reference, a high-precision potential divider, very low signal noise, and very high conversion accuracy of the ADC. This may not be practically achievable in many imaging systems. Furthermore, Vref may be very small, e.g., 1.2 V, which means that a large number of unique levels may need to be used within a relatively small voltage range, further compounding the difficulty.

[0027] FIG. 5 illustrates an exemplary implementation of an imaging system 500 according to the present disclosure. In this example, rather than having a single column of diagnostic pixels, multiple columns of diagnostic pixels are used (four reference columns in this example, Column_D1 through Column_D4, although any suitable number, such as two, three, five, or six columns, could alternatively be present). The reference voltages for each diagnostic pixel can be set so that the combination of reference voltages for each row is different. For example, if only two reference voltage levels are available (i.e., a high voltage such as 1 V and a low voltage such as 0.2 V), it can be seen that in this example of four columns of diagnostic pixels, each row of diagnostic pixels can effectively represent a four-bit digital word. Thus, 16 rows can be uniquely identified using only two voltage references. This number can be increased by adding more columns of diagnostic pixels and / or increasing the number of available voltage reference levels. For example, if eight different voltage levels are available, each diagnostic pixel may effectively generate three bits of ID data (e.g., 000 if it is set to the lowest voltage reference, 001 if it is set to the next lowest voltage reference, 111 if it is set to the highest voltage reference, etc.). Thus, for four columns of diagnostic pixels as shown in Figure 5, the ID for each row may include 12 bits (three bits from each diagnostic pixel in the row) so that 4096 rows can be uniquely identified. This is illustrated in Figure 5 by the use of shading of different types of diagnostic pixels for different reference voltage levels.

[0028] The coupling of the diagnostic pixels to different voltage reference levels is not shown in Figure 5 for clarity, but it will be understood that this can be achieved in any suitable way, for example, by using a voltage reference (e.g., from a bandgap voltage reference) and a potential divider network to set different reference signal levels and then coupling each diagnostic pixel to the appropriate reference signal.

[0029] Table 1 below shows an example of how the reference signal for each diagnostic pixel can be set. In this example, there are four reference levels r0-r3, which means that for an 8-bit row ID, each diagnostic pixel can contribute 2 bits.

[0030] [Table 1]

[0031] The inventors recognize that when three or more different voltage reference levels are available for each diagnostic pixel, variations in offset and / or gain in the reference used by each of the diagnostic pixels can be a source of error in the diagnostic process performed using the readout diagnostic signals. For example, there may be some offset and / or gain variation in the reference setting circuitry across the image sensor 510, such that (for example) the mid-level reference voltage is slightly different for rows at one end of the image sensor 510 compared to rows at the other end, as a result of offset and / or gain in the circuitry that applies the reference levels to the diagnostic pixels. This can result in some errors in identifying the correct row number from the reference signals 348_D1-348_D4.

[0032] This can be solved by having one column of diagnostic pixels always set to either the minimum or maximum reference level. For that column of diagnostic pixels, the rows can alternately receive the maximum and minimum reference signal levels. For the remaining diagnostic pixels, their diagnostic signals can be classified into one of n different reference signal levels by comparing the diagnostic signal read out to a threshold determined based on the diagnostic signal read out from the maximum / minimum diagnostic pixel and to the diagnostic signals read out from adjacent diagnostic pixels (then resulting in finding the maximum and minimum reference levels). This means that any offset and / or gain in the reference signals sent to a pair of rows should be reflected in the maximum and minimum reference signals, which in turn result in appropriate adjustments to the threshold levels used to determine the reference levels set for the remaining diagnostic pixels in those rows.

[0033] For example, the difference between the maximum and minimum reference levels can be defined as a "range." For example, if there are nine different reference levels that can be set for a diagnostic pixel, eight thresholds need to be determined to have a threshold between each adjacent reference level. The read diagnostic signal can then be compared to the thresholds to find the reference level to set for each diagnostic pixel. Each threshold value n can be found as follows: Threshold(n) = (2n+1) × Range / (2 × (Number of Reference Levels - 1)) + Minimum Reference Level

[0034] Taking the example where there are nine reference levels and therefore eight thresholds, the range can be divided by 16 (i.e., 2×(9−1)), and the thresholds can be set at 1 / 16 of the separation between minimum and maximum (i.e., 1 / 16 of the range + minimum reference level), 3 / 16 of the separation between minimum and maximum, 5 / 16 of the separation between minimum and maximum, etc., up to 15 / 16 of the separation between minimum and maximum. In this way, the thresholds define nine reference levels: 0 / 16 (i.e., minimum reference level), 2 / 16, 4 / 16, 6 / 16, etc., up to 16 / 16 (i.e., maximum reference level).

[0035] So if there are eight thresholds that need to be found to define nine reference levels, the minimum threshold n=0 can be found from range / 16 + minimum reference level, the next minimum threshold n=1 can be found from 3 x range / 16 + minimum reference level, etc.

[0036] Table 2 below shows an example of how the reference levels for each diagnostic pixel can be set. In this example, there are nine different reference levels, r0 through r8. The minimum of the voltage reference range is r0 and the maximum is r8 (e.g., r7 can be obtained from a potential divider network where the voltage across the network is r8-r0). The diagnostic pixel column, column_D4, is the column where the diagnostic pixel is set to either the maximum r8 or minimum r0 reference value. In column_D4, each row alternates between the maximum and minimum.

[0037] [Table 2]

[0038] Table 3 shows exemplary digitized values ​​348 output from control and readout circuitry 320 and representing the diagnostic signals read out from each row of diagnostic pixels.

[0039] [Table 3]

[0040] Looking first at rows 0 and 1, the range of 800 can be found by subtracting the minimum signal value read from column D4 of rows 0 and 1 (i.e., 11) from the maximum signal value read from column D4 of rows 0 and 1 (i.e., 811). Thus, the range for this pair of rows is 800. In this case, the minimum comparison threshold n=0 is 800 / 16+11=61. As can be seen, the diagnostic signal levels for the remaining diagnostic pixels, columns D1 through D3, are all below that threshold and therefore can all be classified as being at reference level r0. Other thresholds for n=1 through 6 can be found in the same manner, as needed.

[0041] Now, looking at rows 16 and 17, the range of 799 can be found by subtracting the minimum signal value read from column D4 of rows 16 and 17 (i.e., 92) from the maximum signal value read from column D4 of rows 16 and 17 (i.e., 891). Thus, the range for this pair of rows is 799. In this case, the minimum comparison threshold n=0 is 799 / 16+92=141. The next minimum comparison threshold n=1 is (3×799) / 16+92=241. The remaining comparison thresholds can be found in a similar manner.

[0042] For rows 16 and 17, the diagnostic signal levels for columns D1 and D3 are below the threshold n=0, so they can be classified as being at the reference level r0. Note that these signal levels exceed the threshold n=0 determined for rows 1 and 2, which in this example demonstrates that if a fixed, predetermined threshold were used for all rows, classification errors could occur for at least some of the diagnostic pixels in rows 1 and 2 and / or rows 16 and 17.

[0043] The diagnostic signal levels for column_D2 of rows 16 and 17 are between the n=0 and n=1 thresholds for those rows and can therefore be classified as being at reference level r1. Thus, row 16 can be found to be at r0, r1, r0, r0, giving a row ID of 0000010000. Row 17 can be found to be at r0, r1, r0, r8, giving a row ID of 0000010001.

[0044] Finally, looking at rows 1022 and 1023, the range of 799 can be found by subtracting the minimum signal value (i.e., 66) read from column_D4 of rows 1022 and 1023 from the maximum signal value (i.e., 865) read from column_D4 of rows 1022 and 1023. Thus, the range for this pair of rows is 799. In this case, the minimum comparison threshold n=0 is 799 / 16+66=115. The maximum comparison threshold n=6 is (13×799) / 16+66=715. The remaining comparison thresholds can be found in a similar manner. For rows 1022 and 1023, the diagnostic signal levels for columns_D1 through D3 all exceed threshold n=6, so they can be classified as being at reference level r7. Thus, row 1022 can be found to be at r7, r7, r7, r0, which gives a row ID of 1111111110. Row 1023 can be found to be at r7, r7, r7, r8, which gives a row ID of 1111111111.

[0045] It can thus be seen that the associated comparison threshold to be used for a pair of rows can be determined using the diagnostic signal values ​​from column_D4 for those rows. Each of the diagnostic signals read out from columns_D1 through column_D3 for a pair of rows can then be compared to a threshold to determine the reference level of each diagnostic pixel in that pair of rows, and thus the row ID of that pair of rows. While FIG. 5 shows a single row being read out at a time, some imaging systems can be configured to read out pairs of rows simultaneously (e.g., rows 1 and 2 are selected and read out simultaneously). When a single row is read out at a time, the digital diagnostic signals 348_D1 through 348_D4 for the first row of the pair can be stored, for example, in memory and then accessed for use in determining the row ID after the digital diagnostic signals 348_D1 through 348_D4 for the second row of the pair are read out. This process can be performed within the imaging system, for example, by a diagnostic unit, or by some other processor / system that receives the digital output of control and readout circuitry 320.

[0046] In this example, there are nine reference levels, but the pixels in columns D1-D3 are only set to eight of the reference levels (r0-r7). Using nine reference levels simplifies the setting of threshold levels, so nine reference levels are used in this example even though only eight reference levels are needed for columns D1-D3. This is because in this case, the reference levels are set using a divide-by-eight algorithm, which is a power of two, making the algorithm easier to implement in digital logic. Alternatively, any pixel in columns D1-D3 could be set to any of the reference levels, including the maximum and minimum reference levels (in this example, any of r0-r8). Furthermore, while nine reference levels are used in this example, any number, such as six, seven, eight, or ten, could be used, taking into account the number of diagnostic pixel columns, the precision with which the reference signal can be set for each diagnostic pixel, and the number of rows that need to be uniquely identified.

[0047] 6 shows an example of the former case, in which diagnostic unit 610 receives digital values ​​348_D1-348_D4, determines row IDs as described above, and then outputs the row IDs accompanying imaging data 348_1-348_M for each row. In addition, diagnostic unit 610 may optionally perform any other suitable diagnostic operations, such as by detecting out-of-sequence row IDs or repeated row IDs (e.g., caused by a control signal remaining high). In this case, diagnostic unit 610 may optionally be configured to output fault data whenever a fault is discovered, which in some cases may simply indicate that a fault has been detected and in other cases may indicate the nature of the fault, if diagnostic unit 610 is configured to determine that nature.

[0048] 7 illustrates a further exemplary implementation of an imaging system 700 according to one aspect of the present disclosure, where the diagnostic pixels are arranged in a single row, Row_D, of pixels, rather than a single column. The examples of FIGS. 3-6 may be considered to represent a first plurality of diagnostic pixels arranged in one or more columns. FIG. 7 may be considered to represent a second plurality of diagnostic pixels arranged in a single row.

[0049] The operation of imaging system 700 is very similar to that shown in Figures 3-5. Each of the M diagnostic pixels is set to a different reference 355_1-355_M, each reference derived from a primary reference Vref using a voltage divider network. Row control unit 730 is configured to control the selection and readout of diagnostic signals from the diagnostic pixels using control signal line 335_D. The selection and readout of diagnostic signals can be performed in exactly the same way as the imaging signals readout from the imaging pixels. In this example, the voltage reference set for each diagnostic pixel can serve as a column ID, so that each column can be uniquely identified by reading row_D and comparing the readout diagnostic signals to diagnostic pixel reference values ​​355_1-355_M (or by comparing them to an appropriately set threshold). Thus, any errors in readout lines 342_1-342_M and / or ADC 340 that cause the column readout order to be incorrect can be identified. Furthermore, if any one or more of the readout lines 342_1-342_M is stuck at a particular voltage or is broken, it will also be identified by an incorrect column ID. This is especially true in the example of Figure 7, where the diagnostic pixel column is located on the opposite side of the image sensor 710 from the side where the readout lines 342_1-342_M output the imaging and diagnostic signals, since an error anywhere along that line will then affect the readout diagnostic signal.

[0050] A further purpose of the second plurality of diagnostic pixels is that the readout of the diagnostic signals uses the same ADCs used to read out the imaging signals. As a result, if there is an error in any of one or more of the ADCs, it will be identified by one or more incorrect column IDs appearing in the digital outputs 348_1 through 348_M when a diagnostic pixel is selected and read out.

[0051] FIG. 8 illustrates a further exemplary imaging system 800 in which the second plurality of diagnostic pixels is similar to the implementation of FIG. 6 . In this example, the second plurality of diagnostic pixels are arranged as multiple rows (two rows in this example, but more rows, e.g., three, four, etc., could be used). The combination of reference signals used for each column of the second plurality of diagnostic pixels can be different, in much the same manner as described with reference to FIG. 6 . Thus, by reading out multiple diagnostic pixel rows, many columns (potentially thousands) can be uniquely identified without requiring a high level of diagnostic pixel reference signal precision and / or a low-noise and / or high-precision ADC. Furthermore, because the multiple diagnostic signals read out for each row of diagnostic pixels are different, other errors in the ADC 340 can also be identified. For example, if any of the ADCs is not working, this can be detected because reading out all diagnostic pixel rows forces them to each convert a different reference voltage at some stage during the diagnostic pixel readout. This may require that the reference voltages in each column be set so that they are not all the same, for example, so that at least one diagnostic pixel in each column is set to a reference signal level that is different from the reference signal level of at least one other diagnostic pixel in that column. Thus, if any ADC outputs the same (either identical or similar within a defined tolerance of + / - 1%, + / - 2%, etc.) digital value for each of the diagnostic signals it converts, it may be detected that the ADC is faulty.

[0052] 9 illustrates a further exemplary imaging system 900 according to one aspect of the present disclosure. This example includes both a first plurality of diagnostic pixels and a second plurality of diagnostic pixels, so that both row and column IDs can be determined. This example shows a single column of diagnostic pixels (according to FIGS. 3 and 4) and a single row of diagnostic pixels (according to FIG. 7). Alternatively, it may include multiple columns of diagnostic pixels (according to FIG. 5) and / or multiple rows of diagnostic pixels (according to FIG. 8).

[0053] FIG. 10 shows an exemplary implementation of imaging system 300 in which, rather than coupling each of the diagnostic pixels to a reference signal, each of the diagnostic pixels is an imaging pixel (e.g., of the type depicted in FIG. 1 ), and some are selectively shaded. In this example, the diagnostic pixels may be referred to as “diagnostic imaging pixels” because they are used for diagnosis and are designed to be capable of imaging. In FIG. 10 , the selectively shaded diagnostic pixels are represented by dark shading. As a result of the shading, which forces these diagnostic imaging pixels to be dark or very close to dark, incident light should not cause any charge accumulation (or even a negligible amount of charge accumulation) on the pixel. In contrast, the non-shaded diagnostic imaging pixels, the white diagnostic pixels in FIG. 10 , should accumulate charge in response to incident light.

[0054] 10, where there is only one column of diagnostic imaging pixels, each row is not uniquely identifiable, but some types of errors can be detected by the alternating pattern of shaded and unshaded diagnostic imaging pixels. For example, if pixels are read out one row at a time (e.g., row 1, then row 2, then row 3, etc.), the readout diagnostic signals 342_D, 348_D should alternate between high and low signals. If two consecutive rows are read out with diagnostic signals that are both high or both low, that indicates a row selection / readout error (e.g., row 1 read out is followed by row 3 but not row 2).

[0055] Shading diagnostic pixels, rather than using a reference signal, can also be used in each of the arrangements of Figures 5 through 9. For example, by including multiple columns and / or rows of diagnostic imaging pixels, each diagnostic imaging pixel effectively represents one bit of information (shaded or unshaded), so that each row and / or column can be uniquely identified. This is very similar to Figures 5 and 8. For example, by having 10 columns of diagnostic imaging pixels, 1000 rows of pixels can be uniquely identified using the diagnostic imaging pixels, and / or by having 10 rows of diagnostic imaging pixels, 1000 columns of pixels can be uniquely identified using the diagnostic imaging pixels.

[0056] Using shading for diagnostic imaging pixels may be effective only if the image sensor 310 is imaging a scene with enough light to distinguish between shadowed and unshadowed pixels from the readout diagnostic signals 342_D, 348_D. However, for some applications, such as industrial applications, the light level of the scene may be consistently high such that this alternative implementation may be very effective.

[0057] FIG. 11 illustrates a further exemplary imaging system 1100 according to one aspect of the present disclosure. In this example, the image sensor 1110 does not include any diagnostic pixels. Instead, the imaging system 1100 includes a diagnostic circuit, in this example a replica of a row control unit 1130, to detect faults in the row selection process. The replica of the row control unit 1130 is configured to be a replica of the row control unit 330 and is coupled to the row control lines 335_n to receive row selection control signals placed on those lines by the row control unit 330. The replica of the row control unit 1130 may receive at least some of the same input signals as the row control unit 330 (as can be seen in FIG. 12 below) and is configured to operate in the same manner as the row control unit 330. The replica of the row control unit 1130 is configured to output a value / signal 1140 that can be used to determine whether control of the row selection is being performed properly.

[0058] Figure 12 shows details of one exemplary implementation of row control unit 330 and row control unit replica 1130. Row control unit 330 is the same as that shown in Figure 4 and operates in the same manner. Row control unit replica 1130 comprises a series of cascaded flip-flops 1210_1 to 1210_N of the same type as flip-flops 410_1 to 410_N. The clock input of each of flip-flops 1210_1 to 1210_N is coupled to a respective row control line 335_1 such that a row select signal on row control line 335_n serves to clock flip-flops 1210_1 to 1210_N. The data input of the first flip-flop 1210_1 is coupled to the same select signal as flip-flop 410_1, and the data output of each flip-flop 1210_1 through 1210_N-1 is coupled to the data input of the next flip-flop 1210_2 through 1210_N. As a result, when the signal on row control line 335_1 goes high, the select signal at the data input of flip-flop 1210_1 should be clocked to the data output of flip-flop 1210_1. Thus, if row control unit 330 and row control lines 335_1 through 335_N are all operating correctly, it can be seen that just as the select signal should cascade through flip-flop 410_n, the select signal should similarly cascade through flip-flop 1210_n. As a result, if everything is functioning correctly, value 1140 should go high when the last row N is selected (i.e., when the number of clock signals reaches the sum corresponding to the selection of the last row N). If not, this indicates that there may be a fault somewhere in the row control unit 330 and / or the control line 335_n. Optionally, rather than outputting only one value 1140, the data output of each flip-flop 1210_n may be output such that the progression of the select signal through the cascade may be observed.

[0059] While FIG. 12 shows one particular example implementation of row control unit replica 1130, it will be understood that row control unit 330 may be implemented in many different ways, and row control unit replica 1130 may be implemented in any suitable corresponding way to replicate row control unit 330.

[0060] FIG. 13 illustrates a further exemplary imaging system 1300 according to one aspect of the present disclosure. In this example, the image sensor 1110 is of the same type as in FIGS. 11-13 and does not include any diagnostic pixels. System 1300 is similar to system 1100, but rather than having a replica of row control unit 1130, the diagnostic circuitry is instead diagnostic logic 1310. Diagnostic logic 1310 is configured to process signals on row control lines 335_n in such a way that faults with row control unit 330 and / or row control lines 335_n, such as row control faults or row readout order faults, are detectable from output signal 1320. Diagnostic logic 1310 can be configured in many different ways, depending on the nature of the fault to be detected and the geometry (e.g., size) of image sensor 1110 and the operation of row control unit 330.

[0061] FIG. 14 illustrates one exemplary implementation of the diagnostic logic 1310. For simplicity, the functionality of the diagnostic logic 1310 is depicted for only three rows, rows 1-3, but it will be readily understood that the same functionality can be scaled to all N rows. The diagnostic logic 1310 includes three ID units for the three rows: ID1 1410_1, ID2 1410_2, and ID3 1410_3. ID1 1410_1 is configured to output a value that uniquely indicates row 1, e.g., a binary value of "01." ID2 1410_2 is configured to output a value that uniquely indicates row 2, e.g., a binary value of "10." ID3 1410_3 is configured to output a value that uniquely indicates row 3, e.g., a binary value of "11." It will be understood that if there are four or more rows, longer binary words can be output so that each row can be uniquely identified by a binary word. ID1 1410_1, ID2 1410_2, and ID3 1410_3 may be implemented in any suitable manner, for example, using memory or fixed logic. The diagnostic logic unit also includes AND gates, each coupled to a respective ID unit and a row control line 335_n. As a result, when a row control line goes high, the output from the corresponding AND gate should be the value output by the corresponding ID unit.

[0062] More specifically, the output of each AND gate can be expressed as: row_identifier_gated[n] = row_identifier[n] & row_selection[n] During the ceremony: row_identifier_gated[n] is the output of AND gate n. Row_identifier[n] is the binary value output by ID unit n. Row_select[n] is the signal on row control line 335_n.

[0063] Thus, if row_select[n] is 0, row_identifier_gated[n] is also 0. If row_select[n] is 1, row_identifier_gated[n] is set to row_identifier[n].

[0064] The output of the AND gate is XORd with the combination of the OR gates in Figure 14. As a result, the value of signal 1320 can be expressed as: signal 1320 = row_identifier_gated[1] ∧ row_identifier_gated[2] ∧ row_identifier_gated[3] If imaging system 1300 is operating correctly and only one pixel row is selected at a time, one of the row_identifier_gated signals should be set to the row_identifier value of the selected row, and all other row_identifier_gated signals should be 0. Because 0 does nothing to the XOR gate, the identifier of the selected row should be passed to signal 1320 unchanged. Thus, signal 1320 should indicate the currently selected pixel row, so that selection of the correct pixel row can be verified by signal 1320. As a result, a failure in selecting pixel rows in the wrong order can be identified using signal 1320. A failure (e.g., a break) in row control line 335 can also be identified by signal 1320 never being set to the row identifier value for that row.

[0065] Additionally, signal 1320 can help identify failures where two rows are selected at the same time, because this would corrupt the value set on signal 1320 such that it differs from the expected value.

[0066] For example, if row [2] should be selected, but both row [1] and row [2] are selected at the same time by mistake, row_identifier_gated[1]=2'b01&2'b11=2'b01 row_identifier_gated[2]=2'b10&2'b11=2'b10 row_identifier_gated[3]=2'b11&2'b00=2'b00 Signal 1320=2'b01∧2'b10∧2'b00=2'b11

[0067] Therefore, signal 1320 indicates that row [3] should be selected (because 2'b11 is the row identifier value for row [3]), which is different from the row that should be selected, so it is clear that imaging system 1300 is operating erroneously.

[0068] While FIG. 14 shows one particular logic implementation, it will be understood that a variety of different logic configurations can be used such that the output signal 1320 is generated based on the control signals on the row control lines 335_n in such a way that the output signal 1320 can be used for operational diagnostics (e.g., to determine which row is selected at any given time and whether any row selection faults exist).

[0069] Those skilled in the art will readily appreciate that various changes or modifications can be made to the above-described aspects of the present disclosure without departing from the scope of the present disclosure.

[0070] In each of the above examples, the reference signals used for the diagnostic pixels are derived from a primary voltage reference Vref using a potential divider network. In one exemplary alternative, they may each be generated by a digital-to-analog converter (DAC). This alternative may allow for a crude parametric test of all ADCs in the system, testing, for example, their linearity, gain, full-scale range, etc., by sweeping the DAC through a range of voltages (e.g., the DAC's entire conversion range) with the reference signal being read out at each level the DAC is swept. From the reference signal read out at each level, the ADC's linearity, gain, and / or full-scale range may be determined. Operations such as these may be performed, for example, at system startup and / or intermittently or periodically during system operation (e.g., by simply changing the reference voltage level for each new frame capture using the DAC). Furthermore, it may be possible to change the voltage reference between frames so that the unique row ID and / or column ID changes. This would help detect frames that remain fixed, even though consecutive frames appear identical, only because the image signal level remains fixed. If the reference level for the diagnostic pixel changes from frame to frame and the reference signal output from the ADC remains the same, it indicates a fault in at least part of the imager system causing the fixed frames.

[0071] Although the reference signal applied to and the diagnostic signal read out from each diagnostic pixel are typically described above as voltage signals, in other image sensor implementations they may be any other suitable type of signal. For example, the diagnostic pixels may each include a capacitor, and the reference signal may be a current signal used to charge the capacitor to a reference charge level. In that case, the diagnostic signal may depend on the reference charge level and thus be read out as a voltage signal that depends on the reference current signal.

[0072] Aspects of the Disclosure Non-limiting aspects of this disclosure are set forth in the following numbered clauses:

[0073] Clause 1. An imaging system, comprising: 1. An image sensor, comprising: a plurality of imaging pixels arranged in a matrix of rows and columns, each imaging pixel configured to accumulate charge dependent on light incident on the imaging pixel; an image sensor comprising: at least one diagnostic pixel coupled to a reference signal; A control and readout circuit, selecting one or more of the imaging pixels and, for each selected imaging pixel, reading an imaging signal indicative of the charge accumulated on the selected imaging pixel; selecting at least one diagnostic pixel and reading a diagnostic signal indicative of the reference signal; and a control and readout circuit configured to output a diagnostic signal for use in system diagnostics.

[0074] Clause 2. A first plurality of diagnostic pixels, each coupled to a respective reference signal; the first plurality of diagnostic pixels arranged as one or more columns of diagnostic pixels; A control and readout circuit Selecting and reading out imaging pixels in a specific row of the image sensor; 10. The imaging system of claim 1, configured to select and read out one or more diagnostic pixels in a particular row of the image sensor.

[0075] Clause 3. The imaging system of clause 2, wherein the first plurality of diagnostic pixels are arranged as a single column of diagnostic pixels.

[0076] Clause 4. The imaging system of clause 3, wherein the reference signal for each of the first plurality of diagnostic pixels is different.

[0077] Clause 5. The imaging system of clause 2, wherein the first plurality of diagnostic pixels are arranged as a plurality of columns of diagnostic pixels.

[0078] Clause 6. The imaging system of clause 5, wherein a different combination of reference signals is used for each row of the first plurality of diagnostic pixels.

[0079] Clause 7. The control and readout circuitry comprises a row control circuit configured to output control signals to control the selection and readout of a particular row of imaging pixels and at least one diagnostic pixel; a row control unit positioned on one side of the image sensor adjacent one end column of imaging pixels; 7. The imaging system of any one of clauses 2 to 6, wherein one or more columns of diagnostic pixels are positioned on an opposite side of the image sensor relative to a side of the row control unit.

[0080] Clause 8. An imaging system according to any one of clauses 2 to 7, wherein the control and readout circuitry is configured to simultaneously select and readout a particular row of imaging pixels and one or more diagnostic pixels.

[0081] Clause 9. A second plurality of diagnostic pixels, each coupled to a respective reference signal; the second plurality of diagnostic pixels arranged as one or more rows of diagnostic pixels; A control and readout circuit Selecting one row of imaging pixels and reading out the imaging signals thereof via a plurality of readout lines; 9. The imaging system of claim 1, configured to select one row of one or more rows of diagnostic pixels and read out their diagnostic signals via a plurality of readout lines.

[0082] Clause 10. The imaging system of clause 9, wherein the second plurality of diagnostic pixels is arranged as a single row of diagnostic pixels.

[0083] Clause 11. The imaging system of clause 10, wherein the reference signal for each of the second plurality of diagnostic pixels is different.

[0084] Clause 12. The imaging system of clause 9, wherein the second plurality of diagnostic pixels is arranged as a plurality of rows of diagnostic pixels.

[0085] Clause 13. The imaging system of clause 12, wherein a different combination of reference signals is used for each column of the second plurality of diagnostic pixels.

[0086] Clause 14. An imaging system described in any one of clauses 9 to 13, wherein one or more rows of diagnostic pixels are positioned on the opposite side of the image sensor from the side on which the multiple readout lines output imaging signals and diagnostic signals.

[0087] Clause 15. The device further comprises a diagnostic unit configured to detect faults in the image sensor and / or the control and readout circuitry based on a comparison of the diagnostic signal with a reference signal. 15. The imaging system according to any one of clauses 1 to 14.

[0088] Clause 16. An imaging system, comprising: 1. An image sensor, comprising: a plurality of imaging pixels arranged in a matrix of rows and columns, each imaging pixel configured to accumulate charge dependent on light incident on the imaging pixel; an image sensor comprising, for each row of imaging pixels, a row control line for selecting the row of imaging pixels for readout; a row control circuit coupled to each row control line and configured to output a control signal to each row control line for controlling the selection and readout of the imaging pixels; diagnostic circuitry coupled to each row control line and configured to output a signal generated based on a control signal on the row control line during selection and readout of imaging pixels.

[0089] Clause 17. The imaging system of clause 16, wherein the diagnostic circuitry is a replica of the row control circuitry configured to replicate the operation of the row control circuitry.

[0090] Clause 18. An imaging system as described in clause 16, wherein the diagnostic circuitry comprises logic configured such that during selection and readout of imaging pixels, signals output by the diagnostic circuitry identify pixels of the selected row.

[0091] Clause 19. The imaging system of any one of clauses 16 to 18, wherein a row control circuit is coupled to one end of each row control line, and a diagnostic circuit is coupled to the other end of each row control line.

[0092] Clause 20. An imaging system, comprising: 1. An image sensor, comprising: a plurality of imaging pixels arranged in a matrix of rows and columns, each imaging pixel configured to accumulate charge dependent on light incident on the imaging pixel; an image sensor comprising: a first plurality of diagnostic imaging pixels, some of which are shaded such that they do not accumulate charge dependent on light incident on the diagnostic imaging pixel, and other of which are not shaded such that they accumulate charge dependent on light incident on the diagnostic imaging pixel; A control and readout circuit, selecting one or more of the imaging pixels and, for each selected imaging pixel, reading an imaging signal indicative of the charge accumulated on the selected imaging pixel; selecting at least one of the plurality of diagnostic imaging pixels and, for each selected diagnostic pixel, reading out a diagnostic signal indicative of the charge accumulated on the selected diagnostic pixel; and a control and readout circuit configured to output a diagnostic signal for use in system diagnostics.

[0093] Clause 21. An imaging system as described in Clause 20, wherein the first plurality of diagnostic imaging pixels are arranged as a single column of diagnostic imaging pixels, with shadowed and unshadowed diagnostic imaging pixels alternating along the single column.

[0094] Clause 22. The imaging system of clause 20, wherein the first plurality of diagnostic imaging pixels are arranged as a plurality of columns of diagnostic imaging pixels.

[0095] Clause 23. The imaging system of clause 22, wherein the combination of shadowed and unshadowed diagnostic imaging pixels used for each row of the first plurality of diagnostic imaging pixels is different.

[0096] Clause 24. A second plurality of diagnostic imaging pixels, some of the diagnostic imaging pixels being shaded such that they do not accumulate charge dependent on light incident on the diagnostic imaging pixel, and other of the diagnostic imaging pixels being not shaded such that they accumulate charge dependent on light incident on the diagnostic imaging pixel; further comprising a second plurality of diagnostic imaging pixels, the second plurality of diagnostic imaging pixels arranged as one or more rows of diagnostic imaging pixels; 24. The imaging system according to any one of clauses 21 to 23.

[0097] Clause 25. The imaging system of clause 20, wherein the first plurality of diagnostic imaging pixels are arranged as one or more rows of diagnostic imaging pixels.

Claims

1. 1. An imaging system, comprising:

1. An image sensor, comprising: a plurality of imaging pixels arranged in a matrix of rows and columns, each imaging pixel configured to accumulate charge dependent on light incident on said imaging pixel; an image sensor comprising: a first plurality of diagnostic pixels each coupled to a respective reference signal, the first plurality of diagnostic pixels arranged as a plurality of columns of diagnostic pixels, and for each row of the first plurality of diagnostic pixels, a unique row ID is identified by a different combination of the respective reference signals; A control and readout circuit, selecting one or more of the imaging pixels of a particular row of the image sensor, and reading, for each selected imaging pixel, an imaging signal indicative of the charge accumulated on the selected imaging pixel; selecting the first plurality of diagnostic pixels in the particular row of the image sensor, and reading, for each selected diagnostic pixel, a diagnostic signal indicative of the reference signal for the selected diagnostic pixel; and control and readout circuitry configured to output the diagnostic signal for use in system diagnostics.

2. the control and readout circuitry comprises row control circuitry configured to output control signals to control the selection and readout of the imaging pixels and the first plurality of diagnostic pixels of the particular row; a row control unit positioned on one side of the image sensor adjacent one end column of imaging pixels; The imaging system of claim 1 , wherein the one or more columns of diagnostic pixels are positioned on an opposite side of the image sensor relative to a side of the row control unit.

3. The imaging system of claim 1 , wherein the control and readout circuitry is configured to simultaneously select and readout the imaging pixels and the one or more diagnostic pixels of the particular row.

4. a second plurality of diagnostic pixels each coupled to a respective reference signal; the second plurality of diagnostic pixels being arranged as one or more rows of diagnostic pixels; the control and readout circuitry Selecting a row of imaging pixels and reading out their imaging signals via a plurality of readout lines; 2. The imaging system of claim 1, configured to select a row of the one or more rows of diagnostic pixels and read out their diagnostic signals via the plurality of readout lines.

5. The imaging system of claim 4 , wherein the second plurality of diagnostic pixels is arranged as a single row of diagnostic pixels.

6. The imaging system of claim 5 , wherein the reference signal for each of the second plurality of diagnostic pixels is different.

7. The imaging system of claim 4 , wherein the second plurality of diagnostic pixels is arranged as a plurality of rows of diagnostic pixels.

8. The imaging system of claim 7 , wherein a different combination of reference signals is used for each column of the second plurality of diagnostic pixels.

9. 5. The imaging system of claim 4, wherein the one or more rows of diagnostic pixels are positioned on an opposite side of the image sensor from a side on which the plurality of readout lines output the imaging and diagnostic signals.

10. a diagnostic unit configured to detect faults in the image sensor and / or the control and readout circuitry based on a comparison of the diagnostic signal for each selected diagnostic pixel with the reference signal. The imaging system according to claim 1 .

Images