CN104251742A - Two-point linear and target and environment-based binary non-linear infrared detector heterogeneity correction method - Google Patents

Abstract

The invention discloses a two-point linear and target and environment-based binary non-linear infrared detector heterogeneity correction method. The two-point linear and target and environment-based binary non-linear infrared detector heterogeneity correction method comprises two-point linear heterogeneity correction and target and environment-based binary non-linear heterogeneity correction. The two-point linear heterogeneity correction comprises the following sub-steps of enabling a focal plane and a plane blackbody origin to be aligned through an optical system; calculating a response mean value of every detection element under high and low temperature and response mean values of all detection elements; calculating correction gain and offset of every detection element, and respectively storing inside a lookup table for utilization during correction; correcting the infrared image according to the gain and an offset coefficient in an LUT (Lookup Table). According to the two-point linear and target and environment-based binary non-linear infrared detector heterogeneity correction method, the response characteristics of a focal plane detector are completely described, different requirements on accuracy of resource expenditure, power consumption and application occasions are fully taken into consideration, the imaging quality of an infrared thermal infrared imager can be effectively improved, and the focal plane heterogeneity is reduced.

Description

Based on the binary nonlinear Non-uniformity Correction for IRFPA of two point Linears and target, environment

Technical field

The present invention relates to a kind of non-uniform correction method, particularly relate to a kind of binary nonlinear Non-uniformity Correction for IRFPA based on two point Linears and target, environment.

Background technology

Due to the difference in manufacture process, the spectral response characteristic of each detection unit in un-cooled infrared focal plane array (Focal Plane Array is called for short FPA) detector there are differences, and is called heterogeneity.Heterogeneity shows as in detector output image the fixed pattern noise (Fixed Pattern Noise is called for short FPN) be superimposed upon in scene.This noise has a strong impact on picture quality, is unfavorable for the application such as target detection, identification, needs to be undertaken suppressing even to eliminate by correcting.Due to heterogeneity in time and the change of environmental baseline and drifting about, only carry out before use being once inadequate based on the demarcation of standard blackbody.But in use periodically the demarcation carried out based on black matrix not only needs interruption of work state, and needs extra optics, plant equipment.Method based on scene can utilize the scene in image or target travel to realize Nonuniformity Correction, does not need extras, does not need interruption of work state, thus obtains extensive research and apply.

In recent years, NUC technology obtains very large development, roughly can be divided into two classes, one class is the correcting algorithm based on reference source, as 2 or Supplements technology, they need known reference radiation calibration source, and in systems in practice, in order to reduce the impact of detector individual features drift on calibration result, need periodically to repeat calibration to system, this just adds the complicacy of system accordingly, and have impact on the normal work of system, reduce the reliability of system, but the method is a kind of more ripe Non-uniformity Correction Algorithm, great advantage is exactly the simplicity of algorithm realization, be easy to realize in real-time system.

Another kind of is correcting algorithm based on scene, is the class algorithm utilizing different modern signal theories to set up, as constant statistics, time Hi-pass filter technology, Kalman filtering algorithm etc.; NUC algorithm based on scene utilizes the information in scene to obtain correction parameter, can, eliminating the impact of parameter characteristic drift on NUC performance in varying degrees, be thus the trend of current correcting algorithm development.

Summary of the invention

The object of the invention is to overcome the deficiencies in the prior art, there is provided a kind of can than the response characteristic more comprehensively describing focus planardetector, and take into full account resource overhead, power consumption and the application scenario different demands to precision, effectively can improve the image quality of thermal infrared imager, reduce the binary nonlinear Non-uniformity Correction for IRFPA based on two point Linears and target, environment of focal plane asymmetric.

The object of the invention is to be achieved through the following technical solutions:

Based on the binary nonlinear Non-uniformity Correction for IRFPA of two point Linears and target, environment, it comprises the binary nonlinear nonuniformity correction of two point Linear nonuniformity corrections and based target, environment; Wherein two point Linear nonuniformity corrections comprise following sub-step:

S11: focal plane is aimed at plane black-body resource by optical system, blackbody radiation uniform irradiation on infrared focal plane array, and is full of the whole visual field of focal plane;

S12: control the temperature of blackbody radiation source at TL;

S13: the response y measuring each detection unit in focal plane _ij(φ L), this measured value completes within a time shutter preset, and response is stored in the first storage unit;

S14: repeat step S13, within the time of a large amount of setting, complete a large amount of test datas, multiplicity is 8 ~ 10 times;

S15: calculate the response mean value of each detection unit under TL

S16: the mean value V calculating the response of all detection units _l;

S17: the temperature of blackbody radiation source is set at TH, and TL<TH;

S18: repeat step S13 ~ S16, calculates each detection unit response mean value under temperature TH and the response mean value V of all detection units _h;

S19: the correcting gain G calculating each detection unit _ijwith side-play amount O _ij, be stored in respectively in look-up table (LUT), take for timing;

S110: according to the gain in look-up table (LUT) and side-play amount coefficient, correct infrared image, exports y after obtaining correcting _ij(n);

The binary nonlinear nonuniformity correction of based target, environment comprises following sub-step:

S21: infrared focal plane detector is aimed at plane black-body resource by optical system, blackbody radiation uniform irradiation is on infrared focal plane detector array, and be full of the whole visual field of infrared focal plane detector, and the temperature of blackbody radiation source is controlled on a certain Temperature Scaling point;

S22: under eight scaling point temperature conditions in a temperature range, very to the response of each probe unit of eight groups of infrared focal plane detectors;

S23: the response for all probe units of whole infrared focal plane detector on each calibration temperature point is averaged, obtains its eight average response values at eight Temperature Scaling point places;

S24: the 8 groups of scaling point response datas obtain first two steps and average response carry out least square curve fitting, obtain correction parameter e, g, h of each probe unit, and are preserved by these correction parameters;

S25: finally to needing the view data corrected to bring correction function Z '=eZ into ²+ gZ+h, carries out non-linear non-uniform and corrects output.

The correcting gain G of each detection unit in described step S19 _ijcomputing formula be:

$G_{\mathrm{ij}}=\frac{V_H-V_L}{{\stackrel{\&OverBar;}{y}}_{\mathrm{ij}}\left({\mathrm{\&phi;}}_H\right)-{\stackrel{\&OverBar;}{y}}_{\mathrm{ij}}\left({\mathrm{\&phi;}}_L\right)}.$

The side-play amount O of each detection unit in described step S19 _ijcomputing formula be:

$0_{\mathrm{ij}}=\frac{V_H{y}_{\mathrm{ij}}\left({\mathrm{\&phi;}}_L\right)-V_L{y}_{\mathrm{ij}}\left({\mathrm{\&phi;}}_H\right)}{{\stackrel{\&OverBar;}{y}}_{\mathrm{ij}}\left({\mathrm{\&phi;}}_L\right)-{\stackrel{\_}{y}}_{\mathrm{ij}}\left({\mathrm{\&phi;}}_H\right)}.$

Output in described step S110 is:

y _ij(n)＝G _ij(n)x _ij(n)+O _ij(n)；

X _ijn () is the output before n moment pixel (i, j) Nonuniformity Correction, y _ijn () is the output after n moment pixel (i, j) 2 Nonuniformity Corrections.

The invention has the beneficial effects as follows: the present invention supports two point Linear Corrections and binary nonlinear Nonuniformity Correction two kinds of methods, algorithms selection can be carried out by software merit rating, target temperature and the heteropical influence factor of environment temperature focal plane detector are taken into full account, can than the response characteristic more comprehensively describing focus planardetector, and take into full account resource overhead, power consumption and the application scenario different demands to precision, effectively can improve the image quality of thermal infrared imager, reduce focal plane asymmetric.

Accompanying drawing explanation

Fig. 1 is infrared image element of the present invention response schematic diagram.

Embodiment

Below in conjunction with accompanying drawing, technical scheme of the present invention is described in further detail, but protection scope of the present invention is not limited to the following stated.

Based on the binary nonlinear Non-uniformity Correction for IRFPA of two point Linears and target, environment, it comprises the binary nonlinear nonuniformity correction of two point Linear nonuniformity corrections and based target, environment; Wherein two point Linear nonuniformity corrections comprise following sub-step:

S11: focal plane is aimed at plane black-body resource by optical system, blackbody radiation uniform irradiation on infrared focal plane array, and is full of the whole visual field of focal plane;

S12: control the temperature of blackbody radiation source at TL;

S13: the response y measuring each detection unit in focal plane _ij(φ L), this measured value completes within a time shutter preset, and response is stored in the first storage unit;

S14: repeat step S13, within the time of a large amount of setting, complete a large amount of test datas, multiplicity is 8 ~ 10 times;

S15: calculate the response mean value of each detection unit under TL

S16: the mean value V calculating the response of all detection units _l;

S17: the temperature of blackbody radiation source is set at TH, and TL<TH;

S18: repeat step S13 ~ S16, calculates each detection unit response mean value under temperature TH and the response mean value V of all detection units _h;

S19: the correcting gain G calculating each detection unit _ijwith side-play amount O _ij, be stored in respectively in look-up table (LUT), take for timing;

S110: according to the gain in look-up table (LUT) and side-play amount coefficient, correct infrared image, exports y after obtaining correcting _ij(n).

The response supposing infrared acquisition unit is linear response, as shown in Fig. 1 (a).As can be seen from the figure the basic theories of uncooled ir thermal imaging system respectively detects unit under identical initial conditions, has different output characteristics.The difference of input-output curve intercept reflects the unevenness of the noise current of detector, the difference of rate of curve reflects the unevenness of responsiveness, Nonuniformity Correction is exactly that the response curve of the detector made under identical radiation condition coincides with a curve, set a typical curve for this reason, the response curve of each detection unit is done respectively and rotates and translation transformation, the curve shown in Fig. 1 (b) He Fig. 1 (c) can be obtained, finally make the response curve of detection unit overlap completely.

The binary nonlinear nonuniformity correction of based target, environment comprises following sub-step:

S21: infrared focal plane detector is aimed at plane black-body resource by optical system, blackbody radiation uniform irradiation is on infrared focal plane detector array, and be full of the whole visual field of infrared focal plane detector, and the temperature of blackbody radiation source is controlled on a certain Temperature Scaling point;

S22: under eight scaling point temperature conditions in a temperature range, very to the response of each probe unit of eight groups of infrared focal plane detectors;

S23: the response for all probe units of whole infrared focal plane detector on each calibration temperature point is averaged, obtains its eight average response values at eight Temperature Scaling point places;

S24: the 8 groups of scaling point response datas obtain first two steps and average response carry out least square curve fitting, obtain correction parameter e, g, h of each probe unit, and are preserved by these correction parameters;

S25: finally to needing the view data corrected to bring correction function Z '=eZ into ²+ gZ+h, carries out non-linear non-uniform and corrects output.

The correcting gain G of each detection unit in described step S19 _ijcomputing formula be:

$G_{\mathrm{ij}}=\frac{V_H-V_L}{{\stackrel{\&OverBar;}{y}}_{\mathrm{ij}}\left({\mathrm{\&phi;}}_H\right)-{\stackrel{\&OverBar;}{y}}_{\mathrm{ij}}\left({\mathrm{\&phi;}}_L\right)}.$

The side-play amount O of each detection unit in described step S19 _ijcomputing formula be:

$0_{\mathrm{ij}}=\frac{V_H{y}_{\mathrm{ij}}\left({\mathrm{\&phi;}}_L\right)-V_L{y}_{\mathrm{ij}}\left({\mathrm{\&phi;}}_H\right)}{{\stackrel{\&OverBar;}{y}}_{\mathrm{ij}}\left({\mathrm{\&phi;}}_L\right)-{\stackrel{\_}{y}}_{\mathrm{ij}}\left({\mathrm{\&phi;}}_H\right)}.$

Output in described step S110 is:

y _ij(n)＝G _ij(n)x _ij(n)+O _ij(n)；

X _ijn () is the output before n moment pixel (i, j) Nonuniformity Correction, y _ijn () is the output after n moment pixel (i, j) 2 Nonuniformity Corrections.

Claims (4)

1. based on the binary nonlinear Non-uniformity Correction for IRFPA of two point Linears and target, environment, it is characterized in that: it comprises binary nonlinear nonuniformity correction two steps of two point Linear nonuniformity corrections and based target, environment; Wherein, two point Linear nonuniformity corrections comprise following sub-step:

S11: focal plane is aimed at plane black-body resource by optical system, blackbody radiation uniform irradiation on infrared focal plane array, and is full of the whole visual field of focal plane;

S12: control the temperature of blackbody radiation source at TL;

S13: the response y measuring each detection unit in focal plane _ij(φ _l), this measured value completes within a time shutter preset, and response is stored in the first storage unit;

S14: repeat step S13, within the time of a large amount of setting, complete a large amount of test datas, multiplicity is 8 ~ 10 times;

S15: calculate the response mean value of each detection unit under TL (φ _l);

S16: the mean value V calculating the response of all detection units _l;

S17: the temperature of blackbody radiation source is set at TH, and TL<TH;

S18: repeat step S13 ~ S16, calculates each detection unit response mean value under temperature TH (φ _h) and all response mean value V detecting unit _h;

S19: the correcting gain G calculating each detection unit _ijwith side-play amount O _ij, be stored in respectively in look-up table (LUT), take for timing;

S110: according to the gain in look-up table (LUT) and side-play amount coefficient, correct infrared image, exports y after obtaining correcting _ij(n);

The binary nonlinear nonuniformity correction of based target, environment comprises following sub-step:

S21: infrared focal plane detector is aimed at plane black-body resource by optical system, blackbody radiation uniform irradiation is on infrared focal plane detector array, and be full of the whole visual field of infrared focal plane detector, and the temperature of blackbody radiation source is controlled on a certain Temperature Scaling point;

S22: under eight scaling point temperature conditions in a temperature range, very to the response of each probe unit of eight groups of infrared focal plane detectors;

S23: the response for all probe units of whole infrared focal plane detector on each calibration temperature point is averaged, obtains its eight average response values at eight Temperature Scaling point places;

S24: the 8 groups of scaling point response datas obtain first two steps and average response carry out least square curve fitting, obtain correction parameter e, g, h of each probe unit, and are preserved by these correction parameters;

S25: finally to needing the view data corrected to bring correction function Z '=eZ into ²+ gZ+h, carries out non-linear non-uniform and corrects output.

2. the binary nonlinear Non-uniformity Correction for IRFPA based on two point Linears and target, environment according to claim 1, is characterized in that: the correcting gain G of each detection unit in described step S19 _ijcomputing formula be:

$G_{\mathrm{ij}}=\frac{V_H-V_L}{{\stackrel{\&OverBar;}{y}}_{\mathrm{ij}}\left({\mathrm{\&phi;}}_H\right)-{\stackrel{\&OverBar;}{y}}_{\mathrm{ij}}\left({\mathrm{\&phi;}}_L\right)}.$

3. the binary nonlinear Non-uniformity Correction for IRFPA based on two point Linears and target, environment according to claim 1, is characterized in that: the side-play amount O of each detection unit in described step S19 _ijcomputing formula be:

$O_{\mathrm{ij}}=\frac{V_H{y}_{\mathrm{ij}}\left({\mathrm{\&phi;}}_L\right)-V_L{y}_{\mathrm{ij}}\left({\mathrm{\&phi;}}_H\right)}{{\stackrel{\&OverBar;}{y}}_{\mathrm{ij}}\left({\mathrm{\&phi;}}_L\right)-{\stackrel{\&OverBar;}{y}}_{\mathrm{ij}}\left({\mathrm{\&phi;}}_H\right)}.$

4. the binary nonlinear Non-uniformity Correction for IRFPA based on two point Linears and target, environment according to claim 1, is characterized in that: the output in described step S110 is:

y _ij(n)＝G _ij(n)x _ij(n)+O _ij(n)；

X _ijn () is the output before n moment pixel (i, j) Nonuniformity Correction, y _ijn () is the output after n moment pixel (i, j) 2 Nonuniformity Corrections.