A method for dynamically training serial data of an image sensor

Abstract

The present invention provides a method for dynamically training serial data of an image sensor, which includes the steps of: setting the relevant part of the training word in the parameter register, setting the training control signal to be valid during the idle period of image sensor data transmission or a specific time interval and keeping it valid for more than 2 timing drive clock cycles, the image sensor continuously sending a corresponding number of special training words with asymmetry to the data receiving module according to the training control signal, the data receiving module comparing the received bytes with the special training words or their shifted variants, determining the data shift offset, and then determining the serial-to-parallel adjustment amount and the serial-to-parallel adjustment direction according to the quantization bit number of the image sensor and the data shift offset, and adjusting the serial-to-parallel conversion node to complete the dynamic training. The present invention can realize real-time dynamic training of asynchronous serial image data without affecting the normal imaging of the image sensor, ensure the correct reception of the sensor image data, and improve the reliability and stability of the image sensor during long-term operation.

Description

Technical Field

[0001] The present invention relates to the technical field of image sensors, and particularly to a method for dynamically training serial data of an image sensor. Background Art

[0002] Image sensors are mainly classified into two categories according to their structures: Charge-Coupled Device (CCD) and Complementary Metal Oxide Semiconductor (CMOS) image sensors. CCD image sensors output analog signals and are mostly used in combination with video processors, which convert the analog signals into digital signals; CMOS image sensors integrate an internal mode conversion function and directly output digital signals. The digital signals of image sensors are further divided into serial data and parallel data: Parallel data can be combined with a clock signal for synchronous transmission, and the interface protocol is simple, but the bus width is large and requires several transmission lines with a certain bit width. The high synchronization requirement during high-speed transmission limits the transmission speed; Serial data is asynchronously transmitted, with a small bus width and only requires 2 transmission lines, and the transmission rate is much higher than that of parallel data. However, the serial data transmission requires a certain data synchronization handshake or data training process, resulting in a complex protocol and synchronization process, which brings certain difficulties to engineering implementation.

[0003] With the development of technology, the performance of image sensors has made remarkable progress, with an increasing number of pixels and an increasing working frequency, resulting in an increasingly high data rate of image data. The transmission accuracy of the parallel mode cannot meet the requirements of high-speed data streams, and serial data transmission has been more widely used. Therefore, image training has become an important technology. Since traversal and comparison are required during training and a sufficient number of traversal positions are also needed for training accuracy, the asynchronous image training of complete serial data requires a certain amount of time. Therefore, it is mostly carried out before the start of imaging or during the imaging interval with sufficient time for training completion, and it cannot be implemented for the continuous imaging with only a few clock cycles of interval gaps. However, in actual engineering projects, due to the influence of space environment radiation or electromagnetic interference, etc., the working state of the internal counter of the image sensor will be affected, resulting in problems such as data bit shift in the output image, causing abnormalities in the serial data that has been stably received through image training. At this time, a dynamic training method that can be completed in a short time is needed to correct the received image data in real time to ensure the correct reception of the sensor image data. Summary of the Invention

[0004] Aiming at the problem of serial data bit shift of an image sensor caused by space environment radiation or electromagnetic interference, etc., the present invention proposes an asynchronous serial image data real-time dynamic training method based on a special training word that does not affect the normal imaging of the image sensor.

[0005] To solve the above problems, the present invention adopts the following technical solutions:

[0006] An image sensor serial data dynamic training method, comprising the following steps:

[0007] Step 1: Set the training word related part of the image sensor parameter register through the communication data bus;

[0008] Step 2: Set the training control signal to be valid during the idle period of image sensor data transmission or a specific time interval, and the training control signal lasts for more than 2 timing drive clock cycles. The image sensor continuously sends a corresponding number of special training words with asymmetry to the data receiving module according to the training control signal;

[0009] Step 3: The data receiving module receives at least one byte corresponding to a special training word or its shifted variant, compares the received byte with the special training word or its shifted variant, determines the data shift offset, and then determines the serial-parallel adjustment amount and the serial-parallel adjustment direction according to the image sensor quantization bit number and the data shift offset;

[0010] Step 4: The data receiving module adjusts the serial-parallel conversion node according to the serial-parallel adjustment amount and the serial-parallel adjustment direction to complete the dynamic training of the image sensor serial data.

[0011] The beneficial effects of the present invention are:

[0012] The image sensor serial data dynamic training method proposed by the present invention can solve the problem of image sensor serial data bit shift caused by space environment radiation or electromagnetic interference, etc. This method uses special training words with asymmetry as a reference during the imaging gap such as the idle period of image sensor data transmission or a specific time interval, and finally determines the serial-parallel adjustment amount and the serial-parallel adjustment direction through multiple relevant comparisons. The correction of the serial data bit shift can be completed within 7 clock cycles, and real-time dynamic training of asynchronous serial image data is realized without affecting the normal imaging of the image sensor, ensuring the correct reception of sensor image data, especially improving the reliability and stability of the image sensor during long-term operation, which has important significance in engineering projects. Description of the Drawings

[0013] Figure 1 It is the timing diagram of the normal data reception state after the asynchronous data training is completed;

[0014] Figure 2 It is the timing diagram of the data reception state after a single event upset occurs in the image sensor frequency division counter;

[0015] Figure 3Flow chart of a method for dynamically training serial data of an image sensor according to an embodiment of the present invention;

[0016] Figure 4 Program flow chart of the method for dynamically training serial data of an image sensor according to the specific embodiment of the present invention;

[0017] Figure 5 Timing diagram of the dynamic training state when the image sensor is working normally;

[0018] Figure 6 Timing diagram of the dynamic training state when the internal frequency division counter of the image sensor has a single event upset and the count value decreases;

[0019] Figure 7 is Figure 6 Corresponding normal data reception state timing diagram after training is completed;

[0020] Figure 8 Timing diagram of the dynamic training state when the internal frequency division counter of the image sensor has a single event upset and the count value increases;

[0021] Figure 9 is Figure 8 Corresponding normal data reception state timing diagram after training is completed. Specific embodiment

[0022] The technical solution of the present invention will be described in detail below in conjunction with the accompanying drawings and preferred embodiments.

[0023] During the initialization phase, the image sensor will perform a complete training so that the camera system can correctly receive the serial image data of the sensor. As Figure 1 shown, the image sensor receives an external high-speed clock, generates an internal working clock of the device through an internal frequency division counter (the frequency division multiple is the sensor image quantization bit number), and at the same time, the image sensor outputs serial image data according to the frequency division counter. The camera core processor (such as an FPGA, ISP chip, etc.) will also have a corresponding serial-to-parallel conversion counter. During image training, it will adjust the serial-to-parallel conversion node to synchronize the serial-to-parallel conversion modules of the image sensor and the core processor, so that the image data after serial-to-parallel conversion is an independent and complete byte, and then after internal clock synchronization of the core processor, subsequent operation processing is performed.

[0024] When external disturbances such as space environment radiation or electromagnetic interference occur, it may cause the internal counter of the image sensor or the core processor to have a level flip, resulting in the failure of the synchronization of the serial-to-parallel conversion modules of the image sensor and the core processor, and the linkage between the two is mismatched, which will cause the problem of bit shift in the serial data of the image sensor. As Figure 2As shown, a single-particle phenomenon causes a single-particle upset in the internal frequency divider counter of the image sensor, resulting in a bit shift in data reception.

[0025] In view of the problem of bit shift in the serial data of the image sensor caused by space environment radiation or electromagnetic interference, etc., this embodiment provides a dynamic training method for the serial data of the image sensor, as Figure 3 shown, the method specifically includes the following steps:

[0026] Step 1: Set the training word-related part of the parameter register of the image sensor through a communication data bus (such as an SPI bus, etc.), so that the image sensor outputs a special training word with asymmetry when the training control signal is valid.

[0027] Step 2: Set the training control signal to be valid during the idle period of image sensor data transmission or a specific time interval, and the training control signal lasts for 2 or more timing drive clock cycles. The image sensor continuously sends a corresponding number (2, 3, …) of special training words with asymmetry to the data receiving module according to this training control signal. At this time, if there is a bit shift in the received image data, there will also be a bit shift in the reception of the special training word, but the 2 or more continuously sent special training words can ensure that at least 1 or more special training words or the bytes corresponding to the shifted variants of the special training word can be completely received in the data received by the data receiving module.

[0028] Step 3: The data receiving module compares the received bytes with the special training word or its shifted variant, and can judge the data shift offset. Based on the data shift offset, the quantization bit number of the image sensor, and the multiple comparison results of the bytes with the special training word or its shifted variant, determine the final serial-to-parallel adjustment amount and the serial-to-parallel adjustment direction.

[0029] Step 4: The data receiving module adjusts the serial-to-parallel conversion node according to the serial-to-parallel adjustment amount and the serial-to-parallel adjustment direction, so as to correctly receive the sensor image data, achieving the effect of real-time dynamic training of asynchronous serial image data without affecting the normal imaging of the image sensor.

[0030] In order to realize real-time dynamic training of asynchronous serial image data, the present invention adopts special training words as references. The characteristic of special training words is asymmetry, that is, the training words themselves should not have symmetry. For example, binary numbers 010, 110011, etc. are symmetrical training words. However, symmetric numbers will make it difficult to distinguish the boundaries during serial transmission, and the data cannot be accurately located by search. Therefore, the special training words used in the present invention must be asymmetric. This embodiment provides two methods for constructing special training words: one is the binary code "010011000111······", which is a 01 interleaved format. In this format, the structure of the special training word starts with 1 0 and 1 1, followed by 2 0s and 2 1s, 3 0s and 3 1s, ······, n 0s and n 1s (n is a positive integer), and the data length is determined by the number of quantization bits of the image sensor data; the other is the binary code "101100111000······", which is a 10 interleaved format. In this format, the structure of the special training word starts with 1 1 and 1 0, followed by 2 1s and 2 0s, 3 1s and 3 0s, ······, n 1s and n 0s (n is a positive integer), and the data length is also determined by the number of quantization bits of the image sensor data. Both configurations can ensure that training words above 3 bits are asymmetric. For example, the special training word of a twelve-bit quantized image may be 0x4C7 (type 1) or 0xB38 (type 2), or the variant data of these two data after roller shifting: 0x98E (type 1 shift variant) or 0xCE2 (type 2 shift variant); the special training words of the remaining quantized bit numbers are similar.

[0031] In the present invention, the training control signal lasting for 2 timing drive clock cycles is the smallest achievable solution, which requires the least resources and has the simplest solution. However, for 3 or more timing drive clock cycles, the present invention can also be implemented in a similar manner. For example, in the case where the training control signal lasts for 3 timing drive clock cycles, the data receiving module receives and identifies the bytes corresponding to 2 or 3 consecutive special training words and their shift variants, and then performs corresponding comparison and judgment to finally determine the serial-to-parallel adjustment amount and serial-to-parallel adjustment direction.

[0032] Here, the training control signal is set to have an effective time of 2 timing drive clock cycles as a specific implementation to illustrate the process of the image sensor serial data dynamic training method proposed by the present invention. After setting the training word related part of the image sensor parameter register, as shown in FIG. Figure 4 As shown, the specific process from step 2 to step 4 is as follows:

[0033] a) Determine whether the dynamic training enable is turned on. If the dynamic training enable is not turned on, repeat step a. If the enable is turned on, jump to step b;

[0034] b) Determine whether the image sensor is in a data transmission idle period or a specific time interval. If so, jump to step c; otherwise, repeat step b.

[0035] c) Generate an image sensor training control signal for 2 timing drive clock cycles, and jump to step d.

[0036] d) Judge the first byte corresponding to the image sensor training control signal to determine whether it is a special training word or its shifted variant. After the judgment, jump to step e.

[0037] e) Judge the second byte corresponding to the image sensor training control signal to determine whether it is a special training word or its shifted variant. After the judgment, jump to step f.

[0038] f) Determine whether both the first byte and the second byte are special training words or their shifted variants. If so, jump to step g; otherwise, jump to step n.

[0039] g) Determine whether the data shift offset is 0. If so, jump to step h; otherwise, jump to step i.

[0040] h) The serial-to-parallel adjustment amount is 0, and the serial-to-parallel adjustment direction remains unchanged. Jump to step s.

[0041] i) Determine whether the data shift offset is greater than 1 and less than 6. If so, jump to step j; otherwise, jump to step k.

[0042] j) The serial-to-parallel adjustment amount is the image sensor quantization bit number minus the data shift offset, and the serial-to-parallel adjustment direction is subtraction. Jump to step s.

[0043] k) Determine whether the data shift offset is greater than 7 and less than 11. If so, jump to step l; otherwise, jump to step m.

[0044] l) The serial-to-parallel adjustment amount is the data shift offset, and the serial-to-parallel adjustment direction is addition. Jump to step s.

[0045] m) The serial-to-parallel adjustment amount is 0, and the serial-to-parallel adjustment direction remains unchanged. Jump to step s.

[0046] n) Determine whether the first byte is a special training word or its shifted variant while the second byte is not. If so, jump to step o; otherwise, jump to step p.

[0047] o) The serial-to-parallel adjustment amount is the data shift offset, and the serial-to-parallel adjustment direction is addition. Jump to step s.

[0048] p) Determine whether the second byte is a special training word or its shifted variant while the first byte is not. If so, jump to step q; otherwise, jump to step r.

[0049] q) The parallel - to - serial adjustment amount is the quantization bit number of the image sensor minus the data shift offset, the parallel - to - serial adjustment direction is subtraction, and jump to step s;

[0050] r) The parallel - to - serial adjustment amount is 0, the parallel - to - serial adjustment direction remains unchanged, and jump to step s;

[0051] s) Adjust the parallel - to - serial conversion node according to the parallel - to - serial adjustment amount and the parallel - to - serial adjustment direction, and jump to step t;

[0052] t) Determine whether the parallel - to - serial conversion node is less than 0. If so, jump to step u; otherwise, jump to step v;

[0053] u) Adjust the parallel - to - serial conversion node to the parallel - to - serial conversion node plus the quantization bit number of the image sensor, and jump to step a;

[0054] v) Determine whether the parallel - to - serial conversion node is greater than or equal to the quantization bit number of the image sensor. If so, jump to step w; otherwise, jump to step a;

[0055] w) Adjust the parallel - to - serial conversion node to the parallel - to - serial conversion node minus the quantization bit number of the image sensor, and jump to step a.

[0056] The data shift offset in the present invention refers to the number of bits by which the data is shifted left relative to the special training word on the premise that the data collected by the data receiving module is a special training word or its shifted variant. Taking the special training word 0x98E (this is a shifted variant of type 1 special training word) under 12 - bit quantization as an example, the data shift offsets between it and each variant are shown in the following table.

[0057] Table 1 Corresponding relationship table of data shift offsets

[0058] Serial number Sampling data Data shift offset 1 0x98E 0 2 0x31D 1 3 0x63A 2 4 0xC74 3 5 0x8E9 4 6 0x1D3 5 7 0x3A6 6 8 0x74C 7 9 0xE98 8 10 0xD31 9 11 0xA63 10 12 0x4C7 11

[0059] The dynamic training method for image sensor data is described below under the conditions of normal serial communication and abnormal data bit shift. Assume that the quantization bit number of the image sensor is 12 bits, the special training word is 0x98E, and the non - image - filling redundant data of the image sensor is 0xD56. As Figure 5 shown, it is the timing diagram of the dynamic training state when the image sensor is working normally. During the idle period of image sensor data transmission or a specific time interval, send training control signals for 2 timing - drive clock cycles, then the image sensor sends 2 special training words. After initialization training, the system normally receives data. Then, within 2 cycles when the training control signal is valid after parallel - to - serial conversion, the system receives 0x98E and 0x98E respectively. According to the program flow, the adjustment amount of the FPGA image data receiving counter is 0, that is, no adjustment is made, and the system maintains the current state and continues to work normally. In Figure 5 - Figure 9Among them, Train Pattern is a special training word, Dummy Data is padding data, and the padding data can be set through registers. When the training control signal is valid (pulled high), the detector (i.e., the image sensor) outputs Train Pattern, that is, the special training word. When the training control signal is invalid (pulled low), the detector (i.e., the image sensor) outputs Dummy Data, that is, the padding data. Dummy Data Mix is the mixed padding data, that is, the mixture of two padding data. TrainPatternByte1 Part+Dummy DataPart means that the corresponding received data is the concatenation of the partial bits of the first byte of the special training word and the partial bits of the padding data.

[0060] The image sensor will perform a complete training during the initialization phase so that the camera system can correctly receive the sensor serial image data. However, when the image sensor is abnormal due to space environment radiation or electromagnetic interference, etc. after that, the received image data will have data bit shifting. When external disturbances such as space environment radiation or electromagnetic interference cause the internal counter of the image sensor to have a level flip, the counter value may increase or decrease.

[0061] As Figure 6 shown, it is the dynamic training state timing diagram when the internal frequency division counter of the image sensor has a single event upset and the counter value decreases. External interference causes the internal frequency division counter of the image sensor to change from 4 to 0, resulting in a change in the internal working state of the image sensor, and the output data order and timing change. This causes the data received by the core processor to have bit shifting. It is the mixture of the data of the previous and next two bytes of the image sensor, that is, the aliased data formed by concatenating bits 7 to 0 of the (N + 1)-th byte and bits 11 to 8 of the N-th byte. In this way, after sending the training control signal for 2 timing drive clock cycles during the idle period or a specific time interval of the image sensor data transmission, the two corresponding bytes actually collected are 0x8ED and 0x8E9 respectively. Among them, 0x8ED is obtained by concatenating bits 7 to 0 of the special training word 0x98E + bits 11 to 8 of the padding data 0xD55, and 0x8E9 is obtained by concatenating bits 7 to 0 + bits 11 to 8 of two special training words. The first byte is not the special training word or its shifted variant, and the second byte is the shifted variant of the special training word. The data shift offset is 4. Based on the judgment results of the two bytes, the serial-parallel adjustment amount of the FPGA image data receiving counter is 8, and the adjustment direction is subtraction. Then the FPGA image serial-parallel conversion node is adjusted from 9 to 1, and the dynamic training is completed, and the data bit shifting phenomenon is corrected. As Figure 7 shown.

[0062] As Figure 8As shown in the figure, it is the timing diagram of the dynamic training state when the internal frequency division counter of the image sensor has a single event upset and the count value increases. External interference causes the internal frequency division counter of the image sensor to change from 3 to 7, resulting in a change in the internal working state of the image sensor, and a change in the order and timing of the output data. This causes the data received by the core processor to be misaligned, which is a mixture of the data of the first two bytes before and after the image sensor, that is, the 3rd to 0th bits of the (N + 2)th byte and the 11th to 4th bits of the (N + 1)th byte are spliced into the aliased data. In this way, after sending the training control signal of 2 timing drive clock cycles during the idle period or a specific time interval of the image sensor data transmission, the two corresponding bytes actually collected are 0xE98 and 0x598 respectively. Among them, 0xE98 is obtained by splicing the 3rd to 0th bits + the 11th to 4th bits of two special training words, and 0x598 is obtained by splicing the 3rd to 0th bits of the padding data 0xD55 + the 11th to 4th bits of the special training word 0x98E. The first byte is a shifted variant of the special training word, and the second byte is not a special training word or its shifted variant. The data shift offset is 8. Combining the judgment results of the two bytes, the adjustment amount of the FPGA image data reception counter is 8, and the adjustment direction is addition. Then the FPGA image serial-to-parallel conversion node is adjusted from 9 to 17. Since it is greater than the quantization bit number 12, it is subtracted by 12 to become 5, and the dynamic training is completed, and the data misalignment phenomenon is corrected. As Figure 9 shown.

[0063] A method for dynamically training serial data of an image sensor proposed by the present invention can solve the problem of serial data misalignment of an image sensor caused by space environment radiation or electromagnetic interference, etc. This method uses an asymmetric special training word as a reference during the imaging gap such as the idle period of the image sensor data transmission or a specific time interval. Through multiple correlation comparisons, the serial-to-parallel adjustment amount and the serial-to-parallel adjustment direction are finally determined. The correction and error correction of the serial data misalignment can be completed within 7 clock cycles (2 clock cycles when the training control signal is pulled high + 1 clock cycle for the data acquisition delay of the special training word + 1 clock cycle for data judgment + 1 clock cycle for acquisition position adjustment + 1 clock cycle for application + 1 clock cycle for image data acquisition delay, a total of 7 clock cycles). Without affecting the normal imaging of the image sensor, real-time dynamic training of asynchronous serial image data is realized, ensuring the correct reception of the sensor image data, especially improving the reliability and stability of the image sensor during long-term operation, which has important significance in engineering projects.

[0064] The technical features of the above-described embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as these combinations of technical features do not conflict, they should all be considered as the scope described in this specification.

[0065] The above-described embodiments merely represent several implementation manners of the present invention. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all fall within the protection scope of the present invention. Therefore, the protection scope of the present invention patent shall be subject to the appended claims.

Claims

1. A method for dynamically training serial data of an image sensor, characterized in that, it includes the following steps: Step 1: Set the training word related part of the image sensor parameter register through the communication data bus; Step 2: Set the training control signal to be valid during the idle period of image sensor data transmission or a specific time interval, and the training control signal lasts for more than 2 timing drive clock cycles. The image sensor continuously sends a corresponding number of special training words with asymmetry to the data receiving module according to the training control signal; Step 3: The data receiving module receives at least one byte corresponding to a special training word or its shifted variant, compares the received byte with the special training word or its shifted variant, determines the data shift offset, and then determines the serial-to-parallel adjustment amount and the serial-to-parallel adjustment direction according to the quantization bits of the image sensor and the data shift offset; Step 4: The data receiving module adjusts the serial-to-parallel conversion node according to the serial-to-parallel adjustment amount and the serial-to-parallel adjustment direction to complete the dynamic training of the image sensor serial data.

2. A method for dynamically training serial data of an image sensor according to claim 1, characterized in that, When the special training word adopts the 01 interleaved format, the structure of the special training word starts from 1 zero and 1 one, and then is 2 zeros and 2 ones, 3 zeros and 3 ones, ······, n zeros and n ones in sequence; when the special training word adopts the 10 interleaved format, the structure of the special training word starts from 1 one and 1 zero, and then is 2 ones and 2 zeros, 3 ones and 3 zeros, ······, n ones and n zeros in sequence; where n is a positive integer.

3. A method for dynamically training serial data of an image sensor according to claim 1 or 2, characterized in that, When the training control signal lasts for 2 timing drive clock cycles, the specific processes of steps 2 to 4 are as follows: a) Judge whether the dynamic training enable is turned on. If so, jump to step b; otherwise, repeat step a; b) Judge whether the image sensor is in the idle period of data transmission or a specific time interval. If so, jump to step c; otherwise, repeat step b; c) Generate an image sensor training control signal for 2 timing drive clock cycles, and jump to step d; d) Judge whether the first byte corresponding to the image sensor training control signal is a special training word or its shifted variant. After the judgment, jump to step e; e) Judge whether the second byte corresponding to the image sensor training control signal is a special training word or its shifted variant. After the judgment, jump to step f; f) Judge whether both the first byte and the second byte are special training words or their shifted variants. If so, jump to step g; otherwise, jump to step n; g) Judge whether the data shift offset is 0. If so, jump to step h; otherwise, jump to step i; h) The serial-to-parallel adjustment amount is 0, and the serial-to-parallel adjustment direction remains unchanged. Jump to step s; i) Judge whether the data shift offset is greater than 1 and less than 6. If so, jump to step j; otherwise, jump to step k; j) The serial-parallel adjustment amount is the quantization bit number of the image sensor minus the data shift offset, the serial-parallel adjustment direction is subtraction, and jump to step s; k) Determine whether the data shift offset is greater than 7 and less than 11. If so, jump to step l; otherwise, jump to step m; l) The serial-parallel adjustment amount is the data shift offset, the serial-parallel adjustment direction is addition, and jump to step s; m) The serial-parallel adjustment amount is 0, the serial-parallel adjustment direction remains unchanged, and jump to step s; n) Determine whether the first byte is a special training word or its shifted variant while the second byte is not. If so, jump to step o; otherwise, jump to step p; o) The serial-parallel adjustment amount is the data shift offset, the serial-parallel adjustment direction is addition, and jump to step s; p) Determine whether the second byte is a training word or its shifted variant while the first byte is not. If so, jump to step q; otherwise, jump to step r; q) The serial-parallel adjustment amount is the quantization bit number of the image sensor minus the data shift offset, the serial-parallel adjustment direction is subtraction, and jump to step s; r) The serial-parallel adjustment amount is 0, the serial-parallel adjustment direction remains unchanged, and jump to step s; s) Adjust the serial-parallel conversion node according to the serial-parallel adjustment amount and the serial-parallel adjustment direction, and jump to step t; t) Determine whether the serial-parallel conversion node is less than 0. If so, jump to step u; otherwise, jump to step v; u) Adjust the serial-parallel conversion node to the serial-parallel conversion node plus the quantization bit number of the image sensor, and jump to step a; v) Determine whether the serial-parallel conversion node is greater than or equal to the quantization bit number of the image sensor. If so, jump to step w; otherwise, jump to step a; w) Adjust the serial-parallel conversion node to the serial-parallel conversion node minus the quantization bit number of the image sensor, and jump to step a.

Images