A multi-parameter joint optimization method for space infrared detection payload

Abstract

The present invention discloses a multi-parameter joint optimization method for a space infrared detection payload, which is used to analyze the detection capability of a space-based infrared detection system for air and space targets. First, the basic parameters of the space-based detection system are input, and the optimal integration time of the system is analyzed by the motion characteristics of the target and the radiation characteristics of the typical background. Then, based on the full-link transmission process of the background radiation in the detection system, the detection sensitivity of the system is inverted and calculated. Then, based on the mechanism of introducing additional system noise due to platform jitter during payload operation, the size of the additional noise on orbit is analyzed using the inter-frame offset and spatial clutter. Finally, starting from the principle of target detection, considering the multi-frame superposition noise suppression and the image frame difference noise superposition effect, the detection capability of the infrared detection system during on-orbit operation is evaluated in combination with the camera's own sensitivity and the additional noise of platform jitter. This method decouples the various parameters of the system, analyzes the detection capability of the system based on the detection mechanism, and is suitable for any camera in operation or development.

Description

Technical Field

[0001] The present invention relates to a multi-parameter joint optimization method for space infrared detection payload, which can be used to analyze and evaluate the target detection capability of an infrared system when it is in orbit. The method decouples various system parameters and analyzes the system detection capability based on the detection mechanism, and is applicable to any camera that is already in orbit or under development. Background Art

[0002] With the increasing maturity of infrared detection technology, the application of infrared detection systems is becoming more and more extensive, especially in the field of space-based infrared detection, which has become an important direction for countries to compete for development. At present, there are various design ideas and methods for domestic infrared detection systems, and the performance of systems in orbit also varies significantly. Scientific detection performance evaluation methods play an important role in reasonable system parameter design, system performance estimation during development, and on-orbit system performance evaluation.

[0003] There are currently two approaches to evaluating system performance. One is to develop detection capability test equipment, such as inventions CN202010112790 and CN201310244263, and the other is to evaluate system performance through modeling and simulation analysis, such as invention CN201911314217. Since the detection capability test equipment can only be used for ground testing, it can only perform a relatively effective analysis of the detection capability of the ground system. However, the detection system running in orbit will be interfered by factors such as platform stability and space clutter, resulting in reduced detection efficiency. Therefore, there is no way to accurately evaluate the detection efficiency of the system running in orbit using ground detection capability test equipment. Invention CN201911314217 essentially still needs to obtain the noise performance of the system through ground equipment testing, and does not take into account the changes in detection efficiency brought about by the increase in system noise after running in orbit.

[0004] Starting from the mechanism of noise generation, the present invention comprehensively analyzes the influencing factors of space payload noise, and refines the noise model to each basic parameter of the detection system, ensuring that the detection performance of the system after on-orbit operation can be predicted during the system parameter design stage. Summary of the invention

[0005] The present invention proposes a multi-parameter joint optimization method for space infrared detection payload. As described in the background technology, the existing detection effectiveness evaluation method has the problems of high test system construction cost and insufficient consideration of factors affecting detection effectiveness. However, the present invention starts from the working mechanism of the detector and adopts a multi-parameter decoupling method to achieve accurate evaluation of the system detection effectiveness.

[0006] The technical solution of the present invention comprises the following steps:

[0007] (1) Input of space-based detection system parameters, including spectrum, aperture, resolution, integration time, integration capacitance, payload boresight pointing stability, threshold signal-to-noise ratio, and energy concentration.

[0008] (2) Calculation of the optimal integration time of the system: taking into account the factors of target motion and strong reflected background radiation, the optimal integration time is optimized with the two constraints that the relative motion between the target and the system does not exceed one pixel during the integration time and the number of electrons in the background radiation response is not saturated.

[0009] (3) Characterization of the infrared payload’s own detection sensitivity;

[0010] 3-1) The camera sensitivity is characterized by noise equivalent radiation intensity, noise equivalent temperature difference or noise equivalent radiance, and normalized to the target feature representation method to facilitate the analysis of noise level;

[0011] 3-2) The detection sensitivity of the detection system itself is characterized by the typical radiance of the earth's background.

[0012] (4) Analysis of additional noise caused by jitter of space detection platform: the additional noise introduced after on-orbit operation is obtained through inter-frame offset of visual axis and space clutter. The calculation formula is:

[0013]

[0014] Among them, VEI and CEI are the equivalent radiation intensity of additional noise and the equivalent radiation intensity of spatial clutter, respectively, and x and y are the offsets in the X and Y directions, respectively, that is, the ratio of the pixel offset distance to the pixel center distance.

[0015] (5) Analysis of the on-orbit detection capability of the payload; starting from the background image multi-frame superposition noise suppression and frame difference detection noise superposition factors, the system detection capability I is established. det The expression of I det The calculation formula is:

[0016]

[0017] Among them I det Indicates the minimum target radiation intensity that the system can detect, SNR th is the target detectable threshold signal-to-noise ratio, EE is the energy concentration, ε tar and ε back are the target and background noise suppression percentages respectively, and NEI is the noise equivalent radiant intensity of the camera itself, that is, the sensitivity of the detection system. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a flow chart of the multi-parameter joint optimization method of the space infrared detection payload of the present invention;

[0019] Figure 2 It is a schematic diagram of different types of earth background spectrum radiation curves;

[0020] Figure 3 This is a schematic diagram of the principle of additional noise introduced due to line of sight jitter when space payloads are running. DETAILED DESCRIPTION

[0021] The method described in the present invention is divided into four parts, namely, detection spectrum rationality analysis, camera sensitivity analysis, platform additional noise analysis and detection efficiency satisfaction analysis. The present invention is further described below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments and example data described herein are only used to explain and facilitate the understanding of the present invention, and are not limited to the present invention.

[0022] First, input the parameters of the space-based detection system, including the spectrum (Δλ:λ1~λ2), aperture (D), and equivalent entrance pupil area (A opt ), optical efficiency (τ opt ), angular resolution (ifov), number of electrons in the full well (N full ), quantum efficiency (η), camera line of sight pointing stability (P s ), threshold signal-to-noise ratio (SNR th ), energy concentration (EE), detection distance (R), etc. The embodiments of this paper use parameters of a geosynchronous orbit infrared detection system for calculation.

[0023] 1) Optimal integration time analysis

[0024] Integration time is the main factor that determines the sensitivity of the detection system. Therefore, before analyzing the sensitivity, first calculate the optimal integration time that the system should use. The integration time is generally limited by two factors, including the relative movement speed between the target and the detection system and the background radiance of the earth, which are called the speed limit and the background limit respectively. The speed limit requires that within the integration time, the relative movement distance of the target cannot exceed the ground resolution GR, GR = R·ifov. The background limit requires that the detector response cannot be saturated within the integration time. It is limited by the typical value of the strong background radiance of the earth. In order to ensure the detection dynamic range of the system, it is generally limited to 50% to 80% of the full well number of electrons. That is, the optimal integration time T of the system int Should be

[0025]

[0026] Among them, L q,back is the photon number of the Earth background radiance in the spectrum band λ1~λ2, the full well limit coefficient α (50%≤α≤80%), v tar Indicates the typical movement speed of the target. The typical speed of a high-speed flying target is usually 1 to 10 Ma.

[0027] In space-based detection, the earth's background is relatively complex and is affected by many factors, such as the type of land objects, land altitude, atmospheric transmittance fluctuations, and solar reflection angle. Its spectral radiance distribution is as follows: Figure 2 As shown in Figure 2. According to the difference in background radiance, it can be divided into high-level cloud strong reflection background and conventional background such as ocean, land, and low-altitude clouds. The two have obvious differences in radiance. For this embodiment, in the Δλ spectrum, the high-level cloud strong reflection background radiance can reach 2.5×10 -6 W / sr / cm 2 , which is converted into the form of photon number: 3.6×10 13 e - / s / sr / cm 2 , the typical radiance of a normal background is 7×10 -7 W / sr / cm 2 , converted into the form of photon number is 1.0×10 13 e - / s / sr / cm 2 The integration time needs to be limited by the strong reflection background radiance limit, α is taken as 75%, and the background limit integration time can be calculated as 156ms by substituting it into formula (1). Taking a high-speed flying target as an example, the typical speed is 10Ma, the ground GR is 540m, and the speed limit integration time is 159ms. Therefore, the optimal integration time of the system is 156ms.

[0028] 2) System sensitivity analysis

[0029] System sensitivity is a representation of the noise of the space payload itself. The noise of the space infrared payload itself comes from the readout noise of the detector. read The quantum noise generated by the accumulation of photoelectrons in the integrating capacitor during scene detection. The number of electrons N that the system responds to for a typical detection background back for

[0030] N back =L q,back ·ifov 2 ·A opt ·τ opt ·η·T int (2)

[0031] Therefore, the noise of the system can be expressed as For different application requirements, the detection sensitivity of the system is expressed in different ways, including noise equivalent temperature difference (NETD), noise equivalent radiant intensity (NEI), noise equivalent radiance (NER), etc. Among them, NETD and NER are mainly for surface targets. The expression of NETD is

[0032]

[0033] Where N back is the number of electrons that the system responds to the typical detection background, n read is the read noise, which is related to the number of electrons in the full well by n read =N full / DR, where DR is the dynamic range of the detector. Δ is the temperature differential contrast, NER Expression

[0034]

[0035] Where h and c represent Planck constant and speed of light respectively. NEI is mainly aimed at surface targets. The expression of NEI is:

[0036]

[0037] The difference between NETD, NEI and NER is that they use different physical quantities to express noise. In essence, they are the same and can be converted into each other.

[0038] For this embodiment, the target is a point target under this detection system, so NEI is used to characterize the sensitivity of the camera. For the calculation of sensitivity, since more than 90% of the earth's background is a normal background, the camera sensitivity is calculated based on the normal background radiance, and the result is 5.1W / sr when substituted into formula (5).

[0039] 3) Analysis of platform additional noise

[0040] Through step (2), the noise performance of the system under ground conditions can be analyzed. When the system is in orbit, the jitter of the satellite platform causes the spatial clutter generated by spatial non-uniformity to be introduced into the noise of the system, which is called platform-added noise. The platform-added noise and the camera's own noise together constitute the system's comprehensive noise during in-orbit operation.

[0041] During in-orbit operation, the platform has a certain amount of random jitter, which causes the detection system's line of sight to randomly shift. The random jitter of the detection system's line of sight can be obtained by the angle between the real-time satellite line of sight and the ideal standard line of sight, which can be obtained by statistical analysis of multiple consecutive observations. Generally, the 3σ probability value or 95% probability value of the mean of the random jitter offset is taken, and the unit is angle (°). This value can be estimated and calculated by combining practical application with design experience, so it can be determined during the system design period.

[0042] The principle of platform additional noise generation is as follows Figure 3As shown in the figure, taking pix1 as the reference pixel, each random jitter will affect the three adjacent pixels around it, namely pix2, pix3 and pix4. Assuming that the offset of the visual axis along the X direction is x and the offset along the Y direction is y, the gray value difference ΔDN caused by random jitter between the two frames is

[0043] ΔDN=DN1-((1-x)(1-y)DN1+x(1-y)DN2+(1-x)yDN3+xyDN4) (6)

[0044] Where x and y are dimensionless, representing the ratio of the angle of the X and Y direction offset to the angular resolution. DN1, DN2, DN3 and DN4 are the grayscale values ​​of pix1, pix2, pix3 and pix4 respectively. Assuming that the grayscale values ​​of each pixel are independent of each other during ground imaging, the fluctuations between the grayscale values ​​of each pixel can be represented by spatial non-uniformity, that is, spatial clutter, using the same unit as the camera sensitivity, that is, clutter equivalent temperature difference CETD, clutter equivalent radiance CER or clutter equivalent radiation intensity CEI. Taking into account the independence of the spatial clutter of each pixel, according to formula (6), the noise equivalent radiation intensity VEI, noise equivalent radiance VER and noise equivalent temperature difference VETD generated by random jitter of the visual axis can be expressed as

[0045]

[0046]

[0047]

[0048] Similarly, VEI, VER and VETD are essentially the same and can be converted into each other.

[0049] For this example, in the Δλ spectrum, the 95% probability value of the Earth background CER is about 1.3×10 -7 W / sr / cm 2 , the CEI is 375W / sr at the current ground resolution. The ratio of the camera’s boresight pointing stability to angular resolution is 0.12, which satisfies When x=y, the VEI value is the maximum. Substituting it into formula (7), the VEI is 40.4W / sr.

[0050] 4) Detection effectiveness analysis

[0051] The analysis of detection performance needs to be combined with a specific target detection algorithm. Different target detection algorithms have different detection mechanisms and different detection performances. For this embodiment, point target detection needs to adopt a detection method based on time domain integrated noise. The expression of the time domain signal-to-noise ratio SNR is:

[0052]

[0053] Among them, N tar Indicates the total number of electrons responding to the target. The signal-to-noise ratio generally refers to the detection signal-to-noise ratio of the central pixel. There is a point diffusion effect after the point target is imaged by the optical system. Therefore, the number of target electrons collected by the central target pixel needs to be multiplied by the system energy concentration EE. orbit represents the comprehensive noise of target detection during orbit operation. Formula (10) is expressed in the form of equivalent radiation intensity as

[0054]

[0055] Among them, I tar is the target radiation intensity at the entrance pupil of the system, and SEI is the comprehensive time domain noise equivalent radiation intensity. During the target time domain detection process, it is necessary to perform frame difference operation on the image, so the actual time domain noise is the effect of superimposing two frames of images. In order to avoid introducing too much comprehensive time domain noise, the time domain noise needs to be suppressed before detection. For background frame images, the time domain noise can be suppressed by superimposing multiple frames. Considering that the noise of each frame image is independent of each other, the time domain noise noise after superimposing multiple frames is mf Compared with the temporal noise of a single frame image sf The relationship can be expressed as

[0056]

[0057] Among them, Num is the number of frames. However, the speed of aerospace targets is extremely fast, and it is impossible to suppress the noise of target images by superimposing multiple frames. Therefore, a single frame image is directly used for target detection.

[0058] Based on the above analysis, SEI can be expressed as

[0059]

[0060] From this, we can deduce the detection capability of the system I det The expression is

[0061] I det =SEI·SNR th / EE (14)

[0062] For this embodiment, it is assumed that 20 frames of background superposition are performed to suppress noise, and NEI, VEI, SNR th ,EE into I det =720.1W / sr, that is, the system can detect targets with a radiation intensity of 956.8W / sr and above.

[0063] So far, the present invention has completed the quantitative evaluation of the detection effectiveness of the space-based infrared system. This method comprehensively considers the noise of the camera itself and the additional noise introduced by the platform jitter, and quantitatively evaluates the detection capability of the infrared system based on the target detection principle.

Claims

1. A multi-parameter joint optimization method for space infrared detection payload, characterized in that: The following steps are involved: (1) Input of space-based detection system parameters, including spectrum, aperture, resolution, integration time, integration capacitance, payload line-of-sight pointing stability, threshold signal-to-noise ratio, and energy concentration; (2) Calculation of the optimal integration time of the system: taking into account the factors of target motion and strong reflected background radiation, the optimal integration time is optimized with the two constraints of the relative motion between the target and the system not exceeding one pixel and the number of electrons in the background radiation response not being saturated during the integration time; (3) Characterization of the infrared payload’s own detection sensitivity; 3-1) The camera sensitivity is characterized by noise equivalent radiation intensity, noise equivalent temperature difference or noise equivalent radiance, and normalized to the target feature representation method to facilitate the analysis of noise level; 3-2) The detection sensitivity of the detection system itself is characterized by the typical radiance of the earth's background; (4) Analysis of additional noise caused by jitter of space detection platform: The additional noise introduced after on-orbit operation is obtained through the inter-frame offset of the visual axis and space clutter. The calculation formula is: Among them, VEI and CEI are the equivalent radiation intensity of additional noise and the equivalent radiation intensity of spatial clutter, respectively, and x and y are the offsets in the X and Y directions, respectively, that is, the ratio of the pixel offset distance to the pixel center distance; (5) Analysis of the on-orbit detection capability of the payload; starting from the background image multi-frame superposition noise suppression and frame difference detection noise superposition factors, the system detection capability I is established. det The expression of I det The calculation formula is: Among them I det Indicates the minimum target radiation intensity that the system can detect, SNR th is the target detectable threshold signal-to-noise ratio, EE is the energy concentration, ε tar and ε back are the target and background noise suppression percentages respectively, and NEI is the noise equivalent radiant intensity of the camera itself, that is, the sensitivity of the detection system.

Images