Distributed temperature control system and method for image sensor

Abstract

The invention discloses a distributed temperature control system for an image sensor, and the system comprises a heat-conducting copper block which is connected with a heat dissipation surface of the image sensor, and a thermoelectric cooler TEC which is disposed at one side, away from the image sensor, of the heat-conducting copper block. A temperature detection array formed by a plurality of temperature sensors is arranged between the heat conduction copper block and the image sensor, the temperature detection range of the temperature detection array covers the whole heat dissipation surface of the image sensor, and the signal input end of the TEC is connected with a temperature control and current driving module. The signal input end of the temperature control and current driving module is connected with the signal output end of the temperature detection array. The temperature of the image sensor is monitored in real time through the temperature detection array, and the temperature control precision is remarkably improved in combination with a dynamic response algorithm based on sliding mode control. The system effectively reduces the problem of uneven temperature distribution when the CMOS image sensor works, reduces the imaging noise caused by dark current, improves the imaging quality, and prolongs the service life of the sensor.

Description

Technical Field

[0001] The present invention relates to the technical field of temperature control systems, and particularly to a distributed temperature control system and method for an image sensor. Background Art

[0002] With the rapid development of fields such as machine vision, medical imaging, aerospace, and remote sensing mapping, higher requirements are put forward for the performance of image acquisition devices. Among them, as the core component of an industrial camera, the working temperature of a CMOS image sensor has an important impact on the imaging effect, service life, and stability. In practical applications, the CMOS image sensor generates heat during long-time exposure or high-frequency imaging processes, resulting in the heating of the sensor focal plane, and then generating dark current noise, which affects the image quality. Currently, the mainstream temperature control solutions mostly adopt single-point temperature detection and cooperate with a thermoelectric cooler (TEC) to adjust the temperature, and the control algorithm is mainly the traditional PID control. However, the PID control is relatively sensitive to system parameter changes and has weak anti-interference ability. Especially when the area of the CMOS sensor gradually increases and the non-uniformity of temperature rise enhances, the traditional solution is difficult to meet the high-precision temperature control requirements. Therefore, developing a temperature regulation system with multi-point monitoring, high-precision control, and fast response speed has become the key direction to improve the performance of image sensors. Summary of the Invention

[0003] The present invention aims to at least solve one of the technical problems in the related art to some extent. For this reason, an object of the present invention is to provide a distributed temperature control system for an image sensor, which can reduce the non-uniformity of the spatial distribution of dark current noise caused by temperature non-uniformity of the image sensor, increase the cooling efficiency of the semiconductor refrigeration system, make the temperature control of the image sensor more accurate, and improve the imaging quality of the image sensor.

[0004] In a first aspect, a distributed temperature control system for an image sensor provided by the present invention includes a heat-conducting copper block connected to the heat dissipation surface of the image sensor and a thermoelectric cooler TEC arranged on the side of the heat-conducting copper block away from the image sensor. A temperature detection array formed by a plurality of temperature sensors is arranged between the heat-conducting copper block and the image sensor, and the temperature detection range of the temperature detection array covers the entire heat dissipation surface of the image sensor. The signal input end of the thermoelectric cooler TEC is connected to a temperature control and current drive module, and the signal input end of the temperature control and current drive module is connected to the signal output end of the temperature detection array. The temperature detection array is distributed in a matrix of three rows and three columns.

[0005] Preferably, a plurality of installation grooves for installing the temperature sensors are formed on the side of the heat-conducting copper block close to the heat dissipation surface of the image sensor, and the temperature detection ends of the temperature sensors are arranged on the side close to the heat dissipation surface of the image sensor.

[0006] In a second aspect, a distributed temperature control method for an image sensor proposed by the present invention includes any one of the above-described distributed temperature control systems for an image sensor. The method steps are as follows:

[0007] S1: Obtain the temperature values fed back by the temperature sensors, and establish a one-to-one correspondence between the temperature values and the positions where the sensors are located to form a temperature detection array output matrix.

[0008] S2: Obtain the maximum temperature value fed back by the sensors in step S1, and determine whether the maximum temperature value fed back by the sensors is greater than a preset temperature threshold. If not, execute step S1; if so, execute step S3.

[0009] S3: Apply the temperature detection array output matrix obtained in step S1, and divide the detection array output matrix into four groups of temperature measurement channels according to the division direction from the center to the four corners.

[0010] S4: Set weight coefficients according to the positions where the sensors are located, and multiply the weight coefficients at the corresponding positions by the temperature values fed back by the temperature sensors to obtain a temperature correction matrix.

[0011] S5: Apply the temperature correction matrix obtained in step S4, and add the corresponding values in each group of temperature measurement channels to obtain a temperature reference value representing the temperature reference value of the i-th channel;

[0012] S6: Apply the temperature detection array output matrix obtained in step S1, and subtract the minimum value from the maximum value in each group of temperature measurement channels to obtain the fluctuation factor δ of the channel i , δ i representing the fluctuation factor of the i-th channel;

[0013] S7: Apply the fluctuation factor δ obtained in step S6 i , the dynamic weight of the i-th channel is:

[0014]

[0015] where σ is a jitter compensation constant, σ > 0;

[0016] S8: Apply the temperature reference value obtained in step S5 and the dynamic weight W i obtained in step S7, and the overall temperature estimate of the image sensor is:

[0017]

[0018] S9: The transfer function of the temperature control object is expressed as follows:

[0019]

[0020] Among them, K is the system gain coefficient, T1 and T2 are the thermal inertia time constants of the heat-conducting copper block and the cold end of the thermoelectric cooler TEC respectively, and s is the sliding mode surface function;

[0021] S10: Apply the overall temperature estimation T of the image sensor obtained in step S8 avg , let the actual temperature T(t) of the image sensor be equal to the overall temperature estimation T avg , take the difference between the actual temperature T(t) of the image sensor and the desired temperature T d (t) as the control target, and define the sliding mode surface function as follows:

[0022] u(t) = λ·e(t) + e(t), e(t) = T d (t) - T(t)

[0023] S11: Set the reaching law:

[0024] u(t) = -η·sign(u(t)) - ρ·u(t)

[0025] S12: Control the input current I(t) of the thermoelectric cooler TEC, dynamically adjust it according to the reaching law u(t) in step S11, to control the power output of the thermoelectric cooler, and the final control law is expressed as:

[0026]

[0027] Among them, η, ρ, and λ are control parameters adjusted according to the system dynamic characteristics, used to adjust the reaching speed, suppress temperature overshoot, and enhance the anti-interference ability, represents the second derivative of T d (t), represents the first derivative of e(t).

[0028] Preferably, in step S1:

[0029] The temperature detection array output matrix of three rows and three columns is:

[0030]

[0031] Among them, the elements in the temperature detection array output matrix are the temperature values detected by the sensors at the corresponding positions.

[0032] Preferably, in step S3, the four groups of temperature measurement channels correspond to:

[0033] Channel 1 in the upper left area: T 00 , T 01 , T 10 , T 11 ;

[0034] Channel 2 in the upper right area: T 01 、T 02 、T 11 、T 12 ;

[0035] Channel 3 in the lower left area: T 10 、T 11 、T 20 、T 21 ;

[0036] Channel 4 in the upper right area: T 10 、T 12 、T 21 、T 22 。

[0037] Preferably, in step S4, the weight coefficient matrix is:

[0038]

[0039] The beneficial effects of the present invention are as follows: The temperature at different spatial positions of the CMOS image sensor is detected by temperature sensors distributed in a matrix, and these temperature data are divided into multiple channels for processing. In each channel, the highest temperature in the center of the temperature data is subtracted from the lowest temperature in the channel to obtain the compensated temperature. Finally, the dynamic weights are obtained. Then, different weights are assigned to the temperatures of these channels according to the working characteristics such as the heat generation characteristics of the sensor, and the final average temperature of the sensor is obtained by weighted averaging. This temperature is input into the established thermoelectric cooling power sliding mode control method to drive the thermoelectric cooler to perform cooling with different powers at different temperatures, realizing precise cooling of the CMOS sensor. This method has better dynamic response and steady-state accuracy. Combining with the distributed temperature detection method, it can well reduce the influence caused by the non-uniform dark current of the CMOS temperature sensor and improve the imaging quality of the CMOS image sensor. Description of the Drawings

[0040] In the drawings:

[0041] Figure 1 is a schematic structural diagram of a distributed temperature control system for an image sensor proposed by the present invention;

[0042] Figure 2 is a schematic diagram of the spatial distribution structure of the heat-conducting copper block and the temperature sensor proposed by the present invention;

[0043] Figure 3 is a logic block diagram of a distributed temperature control system for an image sensor proposed by the present invention. Detailed Embodiments

[0044] Refer to Figure 1, A distributed temperature control system for an image sensor, comprising a heat-conducting copper block connected to the heat dissipation surface of the image sensor and a thermoelectric cooler TEC arranged on the side of the heat-conducting copper block away from the image sensor. A temperature detection array formed by a plurality of temperature sensors is provided between the heat-conducting copper block and the image sensor, and the temperature detection range of the temperature detection array covers the entire heat dissipation surface of the image sensor. The signal input end of the thermoelectric cooler TEC is connected to a temperature control and current drive module, and the signal input end of the temperature control and current drive module is connected to the signal output end of the temperature detection array. The temperature detection array is arranged in a matrix of three rows and three columns.

[0045] Obviously, based on the above, the temperature of the heat dissipation surface of the image sensor can be accurately detected in real time by the provided temperature detection array, and this temperature is input into the temperature control and current drive module.

[0046] In this embodiment, referring to Figure 2 , a plurality of mounting grooves for mounting the temperature sensors are provided on the side of the heat-conducting copper block close to the heat dissipation surface of the image sensor, and the temperature detection ends of the temperature sensors are arranged on the side close to the heat dissipation surface of the image sensor.

[0047] Obviously, based on the above, by arranging the temperature detection ends of the temperature sensors on the side close to the heat dissipation surface of the image sensor, the temperature of the heat dissipation surface of the image sensor can be monitored and detected in real time.

[0048] As another embodiment of the present application, referring to Figure 3 , this embodiment proposes a distributed temperature control method for an image sensor, including any one of the above-mentioned distributed temperature control system solutions for an image sensor. The method steps are as follows:

[0049] S1: Obtain the temperature values fed back by the temperature sensors, and correspond the temperature values to the positions where the sensors are located one by one to form a temperature detection array output matrix;

[0050] In this embodiment:

[0051] The temperature detection array output matrix of three rows and three columns is:

[0052]

[0053] Among them, the elements in the temperature detection array output matrix are the temperature values detected by the sensors at the corresponding positions.

[0054] S2: Obtain the maximum temperature value fed back by the sensors in step S1, determine whether the maximum temperature value fed back by the sensors is greater than a preset temperature threshold. If not, execute step S1; if so, execute step S3;

[0055] S3: Apply the output matrix of the temperature detection array obtained in step S1, and divide the detection array output matrix into four groups of temperature measurement channels according to the division direction from the center to the four corners;

[0056] In this embodiment:

[0057] The four groups of temperature measurement channels correspond to:

[0058] Channel 1 in the upper left area: T 00 、T 01 、T 10 、T 11 ;

[0059] Channel 2 in the upper right area: T 01 、T 02 、T 11 、T 12 ;

[0060] Channel 3 in the lower left area: T 10 、T 11 、T 20 、T 21 ;

[0061] Channel 4 in the upper right area: T 10 、T 12 、T 21 、T 22 .

[0062] S4: Set the weight coefficients according to the positions of the sensors, and multiply the weight coefficients at the corresponding positions by the temperature values fed back by the temperature sensors to obtain the temperature correction matrix;

[0063] In this embodiment:

[0064] The weight coefficient matrix is:

[0065]

[0066] S5: Apply the temperature correction matrix obtained in step S4, and add the corresponding values in each group of temperature measurement channels to obtain the temperature reference value represents the temperature reference value of the i-th channel;

[0067] S6: Apply the output matrix of the temperature detection array obtained in step S1, and subtract the minimum value from the maximum value in each group of temperature measurement channels to obtain the fluctuation factor δ of the channel i , δ i represents the fluctuation factor of the i-th channel;

[0068] S7: Apply the fluctuation factor δ obtained in step S6 i , the dynamic weight of the i-th channel is:

[0069]

[0070] Among them, σ is the dither compensation constant, and σ > 0;

[0071] S8: Apply the temperature reference value obtained in step S5 and the dynamic weight W obtained in step S7 i , and the overall temperature of the image sensor is estimated as:

[0072]

[0073] S9: The transfer function of the temperature control object is expressed as follows:

[0074]

[0075] Among them, K is the system gain coefficient, T1 and T2 are the thermal inertia time constants of the copper heat sink and the cold end of the thermoelectric cooler TEC respectively, and s is the sliding mode surface function;

[0076] S10: Apply the overall temperature estimate T of the image sensor obtained in step S8 avg , let the actual temperature T(t) of the image sensor be equal to the overall temperature estimate T avg , and take the difference between the actual temperature T(t) of the image sensor and the desired temperature T d (t) as the control target, and define the sliding mode surface function as follows:

[0077] u(t) = λ·e(t) + e(t), e(t) = T d (t) - T(t)

[0078] S11: Set the reaching law:

[0079] u(t) = -η·sign(u(t)) - ρ·u(t)

[0080] S12: Control the input current I(t) of the thermoelectric cooler TEC, and dynamically adjust it according to the reaching law u(t) in step S11 to control the power output of the thermoelectric cooler. The final control law is expressed as:

[0081]

[0082] Among them, η, ρ, and λ are control parameters adjusted according to the dynamic characteristics of the system, used to adjust the reaching speed, suppress temperature overshoot, and improve the anti-interference ability. represents the second derivative of T d (t), represents the first derivative of e(t).

Claims

1. A distributed temperature control system for an image sensor, characterized in that: It includes a heat-conducting copper block connected to the heat dissipation surface of the image sensor and a thermoelectric cooler TEC arranged on the side of the heat-conducting copper block away from the image sensor. A temperature detection array formed by multiple temperature sensors is arranged between the heat-conducting copper block and the image sensor, and the temperature detection range of the temperature detection array covers the entire heat dissipation surface of the image sensor. The signal input end of the thermoelectric cooler TEC is connected with a temperature control and current drive module, and the signal input end of the temperature control and current drive module is connected with the signal output end of the temperature detection array. The temperature detection array is distributed in a matrix of three rows and three columns.

2. The distributed temperature control system for an image sensor according to claim 1, wherein: On the side of the heat-conducting copper block close to the heat dissipation surface of the image sensor, a plurality of mounting grooves for mounting the temperature sensors are provided, and the temperature detection ends of the temperature sensors are arranged on the side close to the heat dissipation surface of the image sensor.

3. A distributed temperature control method for an image sensor, characterized in that, It includes the distributed temperature control system for the image sensor according to any one of claims 1-2. The method steps are as follows: S1: Obtain the temperature values fed back by the temperature sensors, and correspond the temperature values to the positions where the sensors are located one by one to form a temperature detection array output matrix. S2: Obtain the maximum temperature value fed back by the sensors in step S1, and determine whether the maximum temperature value fed back by the sensors is greater than a preset temperature threshold. If not, execute step S1; if so, execute step S3. S3: Apply the temperature detection array output matrix obtained in step S1, and divide the detection array output matrix into four groups of temperature measurement channels according to the division direction from the center to the four corners. S4: Set weight coefficients according to the positions where the sensors are located, and multiply the weight coefficients at the corresponding positions by the temperature values fed back by the temperature sensors to obtain a temperature correction matrix. S5: Apply the temperature correction matrix obtained in step S4, and add the corresponding values in each temperature measurement channel to obtain the temperature reference value represents the temperature reference value of the i-th channel; S6: Apply the output matrix of the temperature detection array obtained in step S1, and subtract the minimum value from the maximum value in each temperature measurement channel to obtain the fluctuation factor δ of the channel i , δ i represents the fluctuation factor of the i-th channel; S7: Apply the fluctuation factor δ obtained in step S6 i , and the dynamic weight of the i-th channel is as follows: Where σ is a jitter compensation constant, and σ > 0. S8: Apply the temperature reference value obtained in step S5 and the dynamic weight W obtained in step S7 i , and the overall temperature of the image sensor is estimated as: S9: The transfer function of the temperature control object is expressed as follows: Where K is the system gain coefficient, T1 and T2 are the thermal inertia time constants of the heat-conducting copper block and the cold end of the thermoelectric cooler TEC respectively, and s is the sliding mode surface function. S10: Apply the overall temperature estimate T of the image sensor obtained in step S8 avg , and let the actual temperature T(t) of the image sensor be equal to the overall temperature estimate T avg , take the difference between the actual temperature T(t) of the image sensor and the desired temperature T d (t) as the control target, and define the sliding mode surface function as follows: u(t) = λ·e(t) + e(t), e(t) = T d (t) - T(t) S11: Set the reaching law: u(t) = -η·sign(u(t)) - ρ·u(t) S12: Control the input current i(t) of the thermoelectric cooler TEC, and dynamically adjust it according to the reaching law u(t) in step S11 to control the power output of the thermoelectric cooler. The final control law is expressed as: Among them, η, ρ, and λ are control parameters adjusted according to the dynamic characteristics of the system, used to adjust the approaching speed, suppress temperature overshoot, and enhance the anti-interference ability. Denotes the second derivative of d T(t). Denotes the first derivative of e(t).

4. A distributed temperature control method for an image sensor according to claim 3, characterized in that, In step S1: The temperature detection array output matrix of three rows and three columns is: Where the elements in the temperature detection array output matrix are the temperature values detected by the sensors at the corresponding positions.

5. A distributed temperature control method for an image sensor according to claim 4, characterized in that, In step S3, the four groups of temperature measurement channels correspond to: Channel 1 in the upper left area: T 00 , T 01 , T 10 , T 11 ; Channel 2 in the upper right area: T 01 , T 02 , T 11 , T 12 ; Channel 3 in the lower left area: T 10 、T 11 、T 20 、T 21 ; Channel 4 in the upper right area: T 10 , T 12 , T 21 , T 22 .

6. A distributed temperature control method for an image sensor according to claim 4, characterized in that, In step S4, the weight coefficient matrix is:

Images