CN108111759A - Towards the emulation design method of area array CCD opto-electronic conversion - Google Patents

Abstract

The present invention proposes a simulation design method for area array CCD photoelectric conversion, including: according to the source spectral response curve of the multi-spectral input image, and according to the set simulation band range simulation calculation to obtain the CCD spectral responsivity of the simulation band; establish CCD photoelectric conversion model, and complete the photoelectric conversion with the band irradiance field incident on the detector surface as input, and output the voltage image file of the CCD detector response; establish an analog circuit model, and carry out the analog circuit transmission process of pre-amplification, filtering and post-amplification ; Establish a modulus and quantization model, and perform digital quantization processing on the output voltage value data, complete the quantization process from analog signals to digital signals, and output digital image files. The invention can better simulate the photoelectric conversion process of different loads, is applicable to various full-color imaging loads, and can provide load design parameter references for load designers.

Description

Simulation design method for photoelectric conversion of area array CCD

technical field

The invention belongs to the technical field of satellite remote sensing camera simulation, and relates to a simulation design method for area array CCD photoelectric conversion.

Background technique

Satellite imaging simulation is an important link in the satellite design and pre-research stages, aiming at the research on the quality assurance technology of earth observation imaging under complex imaging conditions. By simulating the influence factors of each link in the satellite imaging process on the image quality, the simulated satellite image can be obtained through simulation, which can support the user's analysis of satellite application capabilities, and then guide the proposal of index requirements; it can also assist in the analysis of satellite imaging quality. Before practical application, through experimental simulation verification, hidden dangers and faults can be diagnosed and eliminated in time, and the quality of development can be effectively improved; in addition, by using simulation technology for research, the number and number of experiments can be greatly reduced, thereby shortening the development cycle, The purpose of saving research and development funds and improving cost performance. For the problems that new high-resolution, hyperspectral, and multi-band earth observation satellites will encounter in the future imaging process, as well as the factors that affect image quality, there are no practical examples to refer to. Therefore, in the early stage of satellite remote sensing camera development, the photoelectric conversion simulation of satellite imaging load is helpful to intuitively and quantitatively analyze the actual status and image quality of satellite operation.

Contents of the invention

The problem to be solved by the invention is to solve the problem that the imaging quality of the camera cannot be accurately and intuitively obtained in the design stage of the remote sensing satellite imaging load.

The technical means to solve the problem is to propose a simulation design method for area array CCD photoelectric conversion, simulate the influencing factors of each stage in the photoelectric conversion process, and provide reference for load design.

The present invention proposes a kind of simulation design method facing area array CCD photoelectric conversion, comprises the following steps:

Step 1, according to the source spectral response curve of multi-spectral input image, obtain the CCD spectral responsivity of simulation band according to the simulation band range simulation of setting;

Step 2, set up the CCD photoelectric conversion model, and complete the photoelectric conversion with the band irradiance field on the incident detector surface as input, and output the voltage image file that the CCD detector responds to;

Step 3, set up the analog circuit model, and carry out the analog circuit transmission process of pre-amplification, filtering and post-amplification to the voltage image file output by the CCD photoelectric conversion model, output voltage value data;

Step 4. Establish an analog quantization model, and perform digital quantization processing on the voltage value data output by the analog circuit model, while considering the quantization noise generated during the quantization process, complete the quantization process from analog signals to digital signals, and output digital image files .

Further, as a kind of preferred technical scheme of the present invention, described step 1 simulation calculation obtains the CCD spectral responsivity of simulation band, comprises:

According to the spectral transmittance in the CCD spectral response curve of the camera optical system and the set simulated band range, the optical system transmittance of the simulated band is extracted;

According to the spectral responsivity curve of the CCD and the set simulation band range, the CCD responsivity extraction of the simulation band is completed.

Further, as a preferred technical solution of the present invention, said step 2 takes the band irradiance field on the incident detector surface as input to complete the photoelectric conversion, specifically:

Input the illuminance image to be sampled, and sample the illuminance image actually received by the CCD;

Calculate the illuminance image actually received by a single detector according to the duty cycle of the CCD components;

According to the responsivity and integration time of a single detector, integrate the single detector;

Calculate and output the voltage of a single probe according to the inconsistency response coefficient of the probe;

According to the influence factors of the CCD dynamic range on the output voltage of a single detector, save and output the voltage image file of the photoelectric conversion of the CCD single detector.

Further, as a preferred technical solution of the present invention, the influence factors of the dynamic range of the CCD on the output voltage of a single probe include dark current, saturation output voltage and noise of the CCD.

Further, as a preferred technical solution of the present invention, said step 3 carries out the analog circuit transmission process of pre-amplification, filtering and post-amplification to the voltage image file, specifically:

Perform impedance matching and amplification on the voltage image file output by the CCD photoelectric conversion model, and output the analog voltage after low-pass filtering according to the set low-pass filter frequency;

Adjust the analog voltage output by the low-pass filter and output the voltage signal according to the set post-amplification gain magnification and bias magnification;

According to the circuit noise generated by the camera noise simulation module, the analog circuit noise is superimposed on the voltage signal output by the post-amplification, and the output voltage value data.

Further, as a preferred technical solution of the present invention, the low-pass filtering in the method includes spatial filtering and frequency filtering.

Further, as a preferred technical solution of the present invention, the frequency filtering adopts a Butterworth low-pass filter.

Further, as a preferred technical solution of the present invention, said step 4 performs digital quantization processing on the voltage value data output by the analog circuit model, specifically:

Superimpose the quantization noise generated by the camera noise simulation module on the voltage value data output by the analog circuit model;

According to the established quantization model, the output voltage value data after the superposition is subjected to modulus quantization, and a digital image file is output.

The effect of the invention is that the simulation design method of the present invention has the following advantages:

(1) On the basis of the existing theory of photoelectric conversion system of remote sensing camera, according to the actual needs of the project task, a model of photoelectric conversion system of remote sensing camera based on detailed device parameters of CCD detector is established, and a spectrum is established according to the characteristics of CCD spectral response curve Response model, and realized the software, expanded the function of photoelectric imaging software, realized the simulation and processing of multi-spectral images; refined the photoelectric imaging process, divided the photoelectric conversion process into four subsystems, and realized the control of each subsystem Simulation studies of the effects on imaging.

(2) By comparing the experimental platform built in the laboratory with the simulated images, the simulation model verification work for the photoelectric conversion system of the remote sensing camera was carried out. The experimental results show that: the MTF decline rate of the image degraded by the simulation software simulation and the experimentally captured image The similarity is 90%, which proves that the simulation software model is correct and reliable.

Description of drawings

Fig. 1 is a schematic flow chart of the simulation design method of the present invention.

Fig. 2 is a graph of the spectral response curve of the present invention.

Fig. 3 is a flow chart of CCD photoelectric conversion modeling and simulation processing in the present invention.

Fig. 4 is a flow chart of the analog circuit modeling simulation process of the present invention.

Fig. 5 is a flow chart of the modular and quantitative modeling and simulation processing of the present invention.

Fig. 6 is a schematic diagram of the resampling process of the present invention.

Fig. 7 is a schematic diagram of sampling area division in the present invention.

Fig. 8 is a working principle diagram of the low-pass filter based on the Fourier transform domain of the present invention.

Detailed ways

Hereinafter, the present invention will be described in detail based on the drawings.

As shown in Figure 1, the present invention proposes a kind of emulation design method facing area array CCD photoelectric conversion, comprises the steps:

Step 1. According to the source spectral response curve of the multi-spectral input image, the CCD spectral responsivity of the simulated band is simulated and calculated according to the set simulated band range.

In described step 1, according to multi-spectral input image source spectral response curve, as shown in Figure 2, according to the spectral response curve of the CCD of camera optics system, and according to simulation band range, analyze the CCD spectral responsivity of simulation band, mainly Proceed as follows:

101) According to the spectral transmittance curve of the camera optical system and the simulation band range, complete the extraction of the optical system transmittance of the simulation band, and save it as a separate file.

102) According to the spectral responsivity curve of CCD and simulation band scope, finish the CCD responsivity extraction of simulation band, save into separate file.

Step 2, establish the CCD photoelectric conversion model, and complete the photoelectric conversion with the band irradiance field incident on the detector surface as input, and output the voltage image file of the CCD detector response.

In the step 2, the principle is as shown in Figure 3. The CCD photoelectric conversion model is established according to the parameters such as the integration time of the CCD and the size of the probe, and the band irradiance field on the incident detector surface is used as input to complete the process. Photoelectric conversion, output the voltage image file of CCD detector response, the main steps are as follows:

201) Input the illuminance image to be sampled, consider the impact of resolution, after the sampling process,

Obtain the illuminance image actually received by the CCD;

202) Calculate the illuminance image actually received by a single probe according to the duty cycle of the CCD components;

203) According to the responsivity and the integration time of a single probe, integral processing is carried out to a single probe;

204) Calculate and then output the voltage Vi of a single probe according to the influence of the inconsistency response coefficient of the probe.

205) Finally, according to the influence of the CCD dynamic range on the output voltage of a single detector, the influence of the dynamic range mainly includes dark current, saturated output voltage and noise of the CCD, and finally save the output voltage image file of the photoelectric conversion output of the CCD single detector.

Step 3, establish an analog circuit model, and carry out the analog circuit transmission process of pre-amplification, filtering and post-amplification to the voltage image file output by the CCD photoelectric conversion model, and output voltage value data.

Its principle is as shown in Figure 4, the analog circuit modeling simulation is used to simulate the voltage image data output by the CCD to carry out analog circuit transmission processes such as pre-amplification, filtering and post-amplification, when simulating each link of the circuit, according to the design circuit The characteristic parameters of the simulated voltage data pass through the response data of these circuit devices. The main steps are as follows:

301) Perform impedance matching and amplification on the analog voltage signal output by the CCD sensor according to the amplification factor of the pre-amplification circuit.

302) According to the 3dB frequency (Hz) of the low-pass filter of the filter circuit, the analog voltage signal output by the pre-amplification circuit is low-pass filtered, and the output analog voltage is output.

303) Adjust the analog voltage output by the filter according to the gain amplification factor and the bias amplification amount of the post amplifier, and output the voltage signal.

304) According to the circuit noise generated by the camera noise simulation module, the analog circuit noise is superimposed on the voltage signal output by the post amplifier circuit, and the voltage value data is output.

Preferably, the low-pass filtering in the method includes spatial filtering and frequency filtering. The frequency filtering adopts Butterworth low-pass filter.

Step 4. Establish an analog quantization model, and perform digital quantization processing on the voltage value data output by the analog circuit model, while considering the quantization noise generated during the quantization process, complete the quantization process from analog signals to digital signals, and output digital image files .

In the step 4, the principle is shown in Figure 5. The analog and quantitative modeling simulation performs digital quantization processing on the voltage value data output by the analog circuit, and at the same time considers the quantization noise generated during the quantization process, and completes the conversion from analog signals to digital signals. the quantification process. , the main steps are as follows:

401) superimposing the quantization noise generated by the camera noise simulation module on the voltage value data output by the analog circuit;

402) Establish a quantization model according to the quantization maximum voltage, minimum voltage and quantization bit number of the analog-to-digital converter;

403) According to the quantization model, carry out modulus quantization on the voltage value data result output after superposition in step 401), complete the quantization process, and output a digital image file.

Through the above photoelectric conversion simulation algorithm of satellite imaging payload, the photoelectric conversion process of different payloads can be better simulated, applicable to various panchromatic imaging payloads, and can provide payload design parameter reference for payload designers.

The specific implementation of the present invention will be further described in detail below with reference to FIGS. 2 to 8 , wherein the CCD photoelectric conversion modeling and simulation processing flow is shown in FIG. 3 . The process flow of analog circuit modeling and simulation is shown in Fig. 4. Figure 5 shows the processing flow of modular and quantitative modeling and simulation.

For spectral response simulation modeling: the responsivity of CCD can be represented by the photocurrent I _L and the number of signal charges N _s accumulated by a pixel with an area A _g during the integration period t, which can be written as follows:

In the formula, Φ is the radiance of the scene. The unit of _SI is expressed in mA/W.

For CCD, the FDA method is often used, so in practical applications, the responsivity can also be defined as the size of the output voltage V _S generated by the unit current density σ of the pixel, namely:

The energy flux density of light radiation is usually expressed by illuminance (lx) in photometry, and can be converted by using the relational formula 1W/m ² =20lx.

For the multi-spectral input image source, the corresponding responsivity can be calculated in different bands according to the division of the spectral bands, assuming that the spectral response curve is shown in Figure 2. The data shown in the figure and the image curve shown are hypothetical values and have no actual physical meaning.

For the input spectrum, when the input wavelength is λ _i , the corresponding spectral responsivity is R _λi , then for the imaging band range λ _i ~ λ _j , i<j, the CCD spectral responsivity is:

The photoelectric conversion process of CCD is mainly the process of converting the optical signal incident on the photosensitive element into an electrical signal. There is also a spatial sampling effect, the main reason is caused by the structure of the CCD area array, it is not continuous, but composed of small facets, these small facets perform spatial sampling on the continuous incident light, in the whole There are three main stages in the photoelectric conversion process of CCD.

The first stage is to convert the incident optical signal into an electrical signal (electron/hole), and collect it through the electric field generated by the electrode;

The second stage is the charge transfer process, which transfers and outputs the electrical signal obtained in the imaging area according to a certain timing control;

The third stage is the signal readout amplification process. Considering the influence factors in the photoelectric conversion process, the photogenerated charges (electrons/holes) are converted into voltage signals and amplified for readout.

(1) CCD resampling. Since the size of the CCD detector cannot be infinitesimally small, it is only possible to perform integral imaging on the regional scene in turn, which is the process of spatial sampling. Considering that both the illuminance image to be sampled and the illuminance image after sampling are discrete digital images, the pixels of the sampled illuminance image can be regarded as rectangular blocks rather than pixel points during sampling. First calculate the coordinate position of each sampling point of the CCD (take the pixel center) on the original illumination map, then calculate the boundary position of each sampled pixel, and re-divide the image pixels, as shown in Figure 6.

Thereafter, the illuminance divided into the same pixel is accumulated. It should be noted that since the CCD pixel size is not necessarily an integer multiple of the original pixel, it is easy to cause the original pixel to be divided by the new pixel boundary. At this time, calculate the illuminance size included in the sampling pixel according to the area percentage entering the new pixel. Here, each sampled new pixel is divided into 9 regions, and the schematic diagram of sampling region division is shown in Figure 7.

The pixels of the illuminance map contained in the middle area are all whole pixels, and their illuminance is directly counted into the sampled pixels. In the corner area, the illuminance is calculated according to the area percentage of the original pixel enclosed in the new pixel, and the length and width of the original pixel matrix in the calculation are all defined as 1, so that the coordinates of B1, B2, B3, and B4 are all integers.

Such as the upper left corner area:

Another example is the upper middle area:

The total illuminance of the sampled pixels surrounded by A1A2A3A4:

E＝E _{upper left} +E _{upper middle} +E _{upper right} +E _left +E _middle +E _right +E _{lower left} +E _{middle lower} +E _{lower right}

(2) Calculation of the actual received illuminance of the detector. During the CCD imaging process, different devices have different duty ratios (mark/space ratio, MSR), and the actual irradiance received by each CCD detector is:

E _real = E × MSR

Note: E _real is the actual received illuminance in the flowchart, and MSR is the DutyRatio (CCD duty cycle) in the parameter list.

(3) Probe integration time. The integration of the probe is the storage process of the photogenerated charge of each photosensitive element separated in space, and the amount of charge stored in the photosensitive element depends on the light intensity incident on the photosensitive element and the integration time. Expressed mathematically as

Among them, g(x, y) is the magnitude of the optical signal received by the CCD at the position of the pixel (x, y), and h(x', y', t) represents any point on the pixel (x, y) ( x', y') The energy density of the optical signal at time t, where T represents the exposure time, and a and b represent the length and width of the pixel.

From the perspective of simulation implementation, for a single-band image, it can be expressed by the following formula.

Where R _λ is the response of the corresponding spectral segment, E _realλ is the irradiance information received by the actual phase surface, and τ is the integration time of the area array CCD.

(4) Probe non-uniformity. In the CCD manufacturing process, the main reasons for the non-uniformity of the pixel response are: the inconsistency of the diffusion concentration of the device substrate, the inconsistency of the probe size, and the inability to accurately control the external extension of the surface transparent electrode in the process, etc. The performance is that under uniform illumination, the amount of charge produced by each detector is different, which eventually leads to different image brightness of each detector.

There is no unified definition of the non-uniformity of the response, but a more rigorous definition method is to use the ratio of the root mean square deviation of the photoresponse to the average value of the response as the photoresponse non-uniformity of the CCD. It is generally considered that the photosensitive element is non-uniform, while the CCD is approximately uniform, which also means that the transfer efficiency is the same every time. Then there are:

here:

is the equivalent voltage of the average original response; m is the total number of digits of the linear CCD; V _on is the equivalent voltage of the original response of the nth photosensitive element.

Due to the existence of the transfer loss, the output signal V _n of the CCD is not equal to the original response V _on of the corresponding photosensitive element. However, according to the previous assumptions, V _on can be indirectly calculated as:

In the formula, N is the number of transfers; n _cp is the phase number of CCD.

In the simulation implementation, a linear function is generally used to characterize this non-uniformity:

y[i]=k[i]×x[i]+b[i]

Among them, x[i] is the voltage value of the i-th probe, y[i] is the output of the corresponding probe, k[i] is the inconsistency response coefficient of the i-th probe; b[i] is the The inconsistency bias coefficient of each probe. In the CCD device manual, photoresponse inconsistency (PRNU) and fixed pattern noise (FPN) are generally given, where b[i] is the inconsistency of spectral response, k[i] is PRNU-FPN, which is the distribution intensity , the distribution of these two parameters obeys the normal distribution, and the standard deviation is mainly determined by FPN and PRNU.

(5) Device noise. Photon noise, shot noise, fat zero noise, transfer noise, dark current noise and output noise mainly exist in the photoelectric conversion process of CCD; analog noise mainly exists in the analog circuit simulation process; quantization noise mainly exists in the analog quantization process . Several main noises exist in CCD photoelectric conversion:

1) Photon noise: The emission process of photons is random. Therefore, the CCD is also considered as a random process when collecting light signal charges. Such a random process constitutes a noise source, which is determined by the nature of photons. of. This kind of noise will be more serious when shooting in low light.

2) Shot noise: The existence of shot noise is due to the discrete and quantized wave packets of light or current in motion. Shot noise occurs when the number of photogenerated charges per unit time is slightly different from the average value. According to the characteristics of white noise, shot noise has nothing to do with frequency, and its power distribution in all frequency ranges is uniform. In an environment with low illumination and inconspicuous contrast, after other noise information is suppressed through circuit processing, shot noise becomes the main noise information in the CCD imaging process, which determines the limit noise level of the CCD device.

3) Fat zero noise: That is to use fat zero charge to supplement the potential well position, so that the signal charge can be transferred through the irregular area, and the charge can be divided into electronic fat zero and optical fat zero. Therefore, the generated noise is also divided into electronic noise and optical noise. The optical noise is mainly determined by the bias light of the CCD, and the electronic noise is determined by the electron injection mechanism.

4) Transfer noise: During the CCD transfer process, the previous charge packet has not been completely transferred, and a part of the charge remains in the potential well, thus forming noise interference to the latter charge packet. There are three main causes of transfer noise: interface state trapping, transfer loss, and bulk state trapping.

5) Dark current noise: The main cause of dark current is that the carriers generated by thermal motion inside the semiconductor are added to the potential well, and then transferred under the action of the driving pulse to output current. Dark current exists even in the absence of light. Dark current can be divided into diffusion dark current and surface dark current.

Diffusion dark current is generated by the free region under the conduction channel and potential well of CCD, and the shorter the diffusion length, the more potential wells, the greater the dark current.

The surface dark current is the current formed by electrons jumping from the interface state to the conduction band under thermal excitation, generating free electrons and then being received by the potential well as dark charges.

All CCD detectors are affected by dark current, which determines the sensitivity and dynamic range of the device. Because the size of the temperature determines the size of the dark current noise generated during the thermal movement, and every time the temperature increases by 5°C to 6°C, the dark current will double to the original.

The dark current is also related to the storage time of the charge packet in the potential well, the longer the storage time, the greater the dark current noise will be. Under weak signal conditions, when the CCD uses long-time integration for imaging, dark current noise will be the main factor affecting the imaging quality.

In addition, in the CCD, the presence of impurities or defects in the local lattice may cause dark current peaks. As the doping concentration increases, the closer the distance to the surface, the greater the electric field intensity. At the closest point to the surface, the electric field strength reaches the maximum, and the dark current peak is more likely to appear. Dark current peaks can cause large fluctuations in the image background.

6) Output noise: The CCD signal is output by converting the signal charge of the CCD into a corresponding voltage through the floating capacitor, and the floating diffusion capacitor is mostly used for output.

Among the above noises, since dark current noise and transfer noise have been considered in the photoelectric conversion section, in the next modeling scheme, the influence of dark current and transfer noise is no longer considered, and shot noise is mainly considered. The output the effect of noise.

For the above noises, different noises should be classified in the simulation process, and different noise models should be established according to different noise types. The noise introduced in the process of CCD sampling and photoelectric conversion is inevitable. The noise introduced by CCD is the main source of system noise, and it is the main factor affecting the system imaging quality. In the simulation implementation, because the noise of CCD is difficult to quantify at the micro level, there is no suitable noise model to analyze these noises one by one, so the quantization model needs to be used in the simulation process to replace the noise at the micro level, and they are equivalent to a White noise superimposed on the end of a circuit.

Since most of the noise generated in the photoelectric conversion process of CCD belongs to Gaussian white noise, the research in this paper and the follow-up simulation will assume that all the noise in the photoelectric conversion process obeys the Gaussian distribution with mean value μ and variance σ:

According to the Gaussian noise model, the Gaussian noise generated by the model is superimposed on the final real signal:

V _real = V + V _noise

The main noises generated in the circuit are dark current noise, output noise and reset noise. Also because there is no suitable noise model to analyze these noises one by one, it is assumed that the noise generated by the circuit module is subject to a normal distribution model, and its probability density function is:

Among them, μ and σ are statistical values of noise, and μ is the average value of the equivalent output noise voltage, which is

σ is the root mean square value of the noise voltage fluctuation with time, that is,

Generally, the analog circuit part of the photoelectric conversion of the imaging load mainly includes four modules: a preamplifier module, a low-pass filter module, an operational amplifier module, and a related double-sampling circuit module. Among them, the main function of the correlated double sampling circuit is to eliminate reset noise, extract useful signals, and transmit them to the analog quantization circuit. In the simulation module, due to too many factors affecting the noise, the noise is simplified to a normal distribution model, and no reset noise information is added, so the correlation double sampling circuit can be ignored when modeling. Therefore, in terms of simulation implementation, the analog circuit can be divided into three parts, that is, the preamplifier circuit, the low-pass filter circuit, and the postamplifier circuit. Analog circuit simulation is to simulate the response of the above-mentioned links to the signal.

According to the characteristic parameters of the circuit, the response data of the simulation signal after passing through these circuit modules.

(1) Pre-amplification circuit and post-amplification circuit. The main function of the preamplifier module is to receive the optical signal and convert the optical signal into an electrical signal, then amplify the electrical signal, and then transmit the amplified electrical signal to the subsequent signal processing system. The preamplifier circuit is very important in the whole process of photoelectric conversion, mainly because the optical signal received by the preamplifier circuit is amplified by the amplifier circuit, the signal is very weak, and it is easy to be submerged in noise, so its performance The quality directly determines the performance of the entire circuit system. In terms of simulation implementation, the entire amplifying circuit can be expressed as the following formula.

V ₁ =V ₀ ×B ₁

In the formula, V ₀ is the initial input voltage image, V ₁ is the voltage after the large circuit is prevented, and B ₁ in the above formula represents the pre-amplification factor.

The post-amplification circuit can be expressed as the following formula in the simulation implementation.

V ₃ =V ₂ ×B ₂ +A

V ₂ is the voltage after low-pass filtering. Multiplying V ₂ with the gain amplification factor B ₂ and adding the bias amplification value A is the final analog circuit simulation output result V ₃ .

(2) Low-pass filter. Generally, there are two common filtering methods: spatial filtering and frequency filtering. Spatial filtering refers to directly changing the space of each pixel of the image in the domain and then processing it, while frequency filtering refers to transforming the image information into the frequency domain after Fourier transformation to process the spectral components. For the two processing methods, spatial domain filtering is a neighborhood operation, while frequency domain filtering is more intuitive. The key issue of frequency domain filtering is to select the filter transfer function, establish a mathematical model, and finally obtain the processed output image by inverse Fourier transforming the image information processed in the frequency domain.

The theoretical basis of spatial domain filtering and frequency domain filtering is the convolution theorem.

The expressions at both ends of the formula are a set of Fourier transform pairs, which means that the convolution of two spatial functions can be obtained by the inverse transform of the product of Fourier transforms. That is to say, in the spatial domain, the convolution operation of the image to be filtered f(x,y) and the filter mask h(x,y) can be multiplied by F(u,v) in the frequency domain by H(u ,v) is calculated.

Filtering in the frequency domain can be seen as using a filter transfer function to modify F(u,v), and then calculating the inverse Fourier transform of H(u,v)F(u,v), which can be obtained accordingly Spatial domain filtered image. Its principle is shown in Figure 8.

The Butterworth filter has the largest flat amplitude response and is widely used in the communication field. Compared with Bessel (bessl) and Chebyshev (chebyshev) filters, Butterworth filter has the advantages of balance in linear phase, attenuation slope and loading characteristics, making Butterworth low-pass filter more suitable for Image Processing.

For an nth-order Butterworth low-pass filter, which has a non-sharp cutoff, its transfer function is:

In the formula, D ₀ is the cut-off frequency, where When D(u,v)=D ₀ , n is the order of the filter, which is a positive integer and is mainly used to control the attenuation speed of the frequency.

Generally speaking, the cut-off frequency of the low-pass filter is generally selected so that H(u,v) drops to its The frequency value at , the transfer function at this time can be written as:

When n=1, the Butterworth low-pass filter has no "ringing" phenomenon, and can improve the contrast of the image; when n=2, there will be a weak "ringing" phenomenon; and when n is larger, The more obvious the "ringing" phenomenon is, the closer it is to an ideal low-pass filter.

Finally, the imaging load modulus and quantization simulation performs digital quantization processing on the voltage value data output by the analog circuit, and at the same time considers the quantization noise generated during the quantization process, and completes the quantization process from the analog signal to the digital signal.

The analog and quantitative simulation performs digital quantization processing on the voltage value data output by the analog circuit, and establishes a quantization model according to the quantized maximum voltage, minimum voltage and quantized digits of the analog-to-digital converter;

Among them, N is the number of digits quantized; V _max and V _min are the output maximum and minimum signal voltages, respectively.

In summary, the present invention has carried out the simulation model verification work for the photoelectric conversion system of the remote sensing camera by building an experimental platform in the laboratory and comparing it with the simulated image. The experimental results show that the MTF of the image degraded by the simulation software simulation and the image taken by the experiment decreases The similarity rate is 90%, which proves that the simulation software model is correct and reliable.

It should be noted that the above description is only a preferred embodiment of the present invention, and it should be understood that for those skilled in the art, several changes and improvements can be made without departing from the technical concept of the present invention, which are included in within the protection scope of the present invention.

Claims (8)

1. a simulation design method for area array CCD photoelectric conversion, it is characterized in that, comprises the steps: Step 1. According to the source spectral response curve of the multi-spectral input image, the CCD spectral responsivity of the simulation band is obtained through simulation calculation according to the set simulation band range; Step 2, establish a CCD photoelectric conversion model, and complete the photoelectric conversion with the band irradiance field on the incident detector surface as input, and output the voltage image file of the CCD detector response; Step 3, establish an analog circuit model, and carry out the analog circuit transmission process of pre-amplification, filtering and post-amplification to the voltage image file output by the CCD photoelectric conversion model, and output voltage value data; Step 4. Establish an analog quantization model, and perform digital quantization processing on the voltage value data output by the analog circuit model, while considering the quantization noise generated during the quantization process, complete the quantization process from analog signals to digital signals, and output digital image files . 2. the simulation design method facing area array CCD photoelectric conversion according to claim 1, is characterized in that, described step 1 simulation calculation obtains the CCD spectral responsivity of simulation band, comprising: According to the spectral transmittance in the CCD spectral response curve of the camera optical system and the set simulated band range, the transmittance of the optical system in the simulated band is extracted; According to the spectral responsivity curve of the CCD and the set simulation band range, the CCD responsivity extraction of the simulation band is completed. 3. the simulation design method facing the photoelectric conversion of the area array CCD according to claim 1, it is characterized in that, described step 2 is input to complete the photoelectric conversion with the band irradiance field on the incident detector face, specifically: Input the illuminance image to be sampled, and sample the illuminance image actually received by the CCD; Calculate the illuminance image actually received by a single detector according to the duty cycle of the CCD components; According to the responsivity and integration time of a single detector, integrate the single detector; Calculate and output the voltage of a single probe according to the inconsistency response coefficient of the probe; According to the influence factors of the CCD dynamic range on the output voltage of a single detector, save and output the voltage image file after the photoelectric conversion of the CCD single detector. 4. the simulation design method facing area array CCD photoelectric conversion according to claim 3, is characterized in that, the influence factor of described CCD dynamic range comprises the noise of dark current, saturated output voltage and CCD to single detector output voltage. 5. the simulation design method facing area array CCD photoelectric conversion according to claim 1, it is characterized in that, described step 3 carries out the analog circuit transmission process of pre-amplification, filtering and post-amplification to voltage image file, specifically: Perform impedance matching and amplification on the voltage image file output by the CCD photoelectric conversion model, and output the analog voltage after low-pass filtering according to the set low-pass filtering frequency; Adjust the analog voltage output by the low-pass filter and output the voltage signal according to the set post-amplification gain magnification and bias magnification; According to the circuit noise generated by the camera noise simulation module, the analog circuit noise is superimposed on the voltage signal output by the post-amplification, and the voltage value data is output. 6. the simulation design method for area array CCD photoelectric conversion according to claim 5, is characterized in that, low-pass filtering comprises spatial filtering and frequency filtering in the described method. 7. The simulation design method for photoelectric conversion of an area array CCD according to claim 6, wherein said frequency filtering adopts a Butterworth low-pass filter. 8. the simulation design method facing area array CCD photoelectric conversion according to claim 1, is characterized in that, described step 4 carries out digital quantization processing to the voltage value data of analog circuit model output, specifically: Superimpose the quantization noise generated by the camera noise simulation module on the voltage value data output by the analog circuit model; According to the established quantization model, the superimposed output voltage value data is subjected to modulus quantization, and a digital image file is output.