Correcting defects caused by dynamic variations in bolometer biasing voltages

Abstract

The invention relates to a method for acquiring and processing images of a scene (4) provided by a detection system (2) comprising a sensor (6), the method comprising: a step of modifying an adjustment parameter of the sensor (6), a step of modifying a reference offset table of the sensor (6), with the modification being applied to the images, and a step of compensating for variations in the average intensity of the images caused by the modifications to the adjustment parameter of the sensor (6) and the reference offset table of the sensor (6), with the compensation being applied to the images.

Description

[0001] DESCRIPTION

[0002] TITLE: CORRECTION OF DEFECTS INDUCED BY DYNAMIC VARIATION OF BOEOMETER POEARIZATION VOLTAGES.

[0003] Technical field

[0004] The technical field of the invention is electromagnetic radiation detection systems, in particular bolometric detection systems for uncooled infrared imaging.

[0005] In particular, the present invention relates to a method and a device for acquiring and processing images.

[0006] Previous techniques

[0007] The temperature of an object is characteristic of its spectral emission. The analysis of electromagnetic radiation coming from a scene containing the object can make it possible to detect the object and characterize its nature. There are applications for detection and / or monitoring and / or optical surveillance of the scene requiring the distinction of objects of similar temperatures and / or requiring the detection of the temperature of objects over a wide temperature range with the same detection system.

[0008] The detection system generally comprises an array of bolometric detectors sensitive to electromagnetic radiation from the scene, as well as a plurality of blind bolometers insensitive to said electromagnetic radiation.

[0009] The detection system also comprises a computer provided with a memory in which adjustment values ​​of the detection system are recorded, for example a value for adjusting the bias voltage of the matrix of bolometric detectors, a value for adjusting the bias voltage of the blind bolometers and a value for adjusting a capacitive transimpedance amplification gain, called "CTIA gain".

[0010] The detection system provides a succession of images of the scene based on measurements from the bolometric detector array, measurements from the blind bolometers and the CTIA gain. The adjustment values ​​of the detection system are sometimes unsuitable for the conditions of implementation of the detection system, harming the quality of the images of the scene formed by the detection system, and therefore the quality of the information acquired on the object to be detected.

[0011] It is known to adapt the adjustment values ​​of the detection system to the observed scene, particularly dynamically. However, the adaptation of these adjustment values ​​causes the appearance of spatial noise in the images of the scene, which disrupts the reading of the images.

[0012] In addition, the setting values ​​are usually discrete values, especially digital ones. Changing the setting values ​​is therefore sometimes done with a large step.

[0013] A significant change in the setting values ​​can cause a significant change in the average intensity value of the scene images.

[0014] In particular, the average pixel intensity value of one image may be significantly higher or significantly lower than the average pixel intensity value of the previous image, thus creating abrupt changes in average intensity in the succession of images.

[0015] To adapt the adjustment values ​​of the detection system to the observed scene, an automatic histogram control algorithm is generally implemented in order to recenter the histogram of the intensity of the images of the scene by correcting the variations in average intensity in the succession of images linked to the natural variations of the scene. However, such recentering of the histogram is done with several images of delay and does not prevent sudden changes in average intensity in the succession of images linked to the adaptation of the adjustment values ​​of the detection system to the observed scene.

[0016] These changes in intensity and the appearance of spatial noise in the images can be annoying for an operator observing the scene using the detection system, degrade the quality of the information acquired on the object to be detected and hinder the operator's decision-making.

[0017] Statement of the invention

[0018] The present invention therefore aims to overcome all or part of the aforementioned drawbacks, to enable the detection of the temperature of objects over a wide range of temperatures, to increase the measurement dynamics of the detection system with respect to the observed scene, in particular by reducing the noise of the detection system, and to avoid sudden variations in intensity between the different images of a succession of images of the scene provided by the detection system.

[0019] Another objective is to make variations in at least one parameter of the detection system transparent to the user by avoiding sudden variations in intensity between the different images in the succession of images of the scene provided by the detection system.

[0020] The present invention relates to a method for acquiring and processing images of a scene provided by a detection system comprising a sensor provided with a matrix of detectors sensitive to electromagnetic radiation coming from the scene, called "active", a plurality of detectors insensitive to said radiation, called "blind", as well as a reading circuit providing a response signal for each sensitive detector, the detection system providing a succession of images of the scene from the response signals, the method comprising: modifying a sensor adjustment parameter; modifying a sensor reference offset table applied to the images; and compensating for a variation in the average intensity of the images generated by the modification of the sensor adjustment parameter and the sensor reference offset table applied to the images.

[0021] Changing the sensor adjustment parameter allows the sensor settings to be adapted to the scene observed by the detection system to avoid saturating hot or cold objects in the observed scene.

[0022] Modifying the sensor reference offset table applied to the images, also called "offset" in English terms, makes it possible to adapt the reference offset applied to each pixel of the scene images provided by the detection system to the new sensor adjustment parameter.

[0023] Compensating for the variation in the average intensity of the images provided by the detection system makes it possible to avoid sudden changes in intensity between two successive images provided by the detection system. Thus, an operator observing the images of the scene provided by the detection system will not see a sudden change in the average intensity level of the pixels of the images of the scene when changing the sensor adjustment parameter and the sensor reference shift table applied to the images. In particular, compensating for the variation in the average intensity of the images is performed to compensate for the variation in intensity resulting from the applied changes in the sensor adjustment parameter and the sensor reference shift table applied to the images.

[0024] Advantageously, the sensor adjustment parameter comprises a bias voltage applied to the insensitive detectors called “base voltage” and / or a value of a trans-impedance capacitive amplification gain of the sensor called “CTIA gain”.

[0025] The sensor tuning parameter may also include a bias voltage applied to the sensitive detectors.

[0026] Modifying the sensor adjustment parameter may comprise calculating an average intensity value of at least a portion of an image of the succession of images, determining a deviation between said average intensity value and a target value, comparing the deviation to a tolerance threshold and, depending on the result of the comparison, applying a new value of the sensor adjustment parameter.

[0027] Alternatively, modifying the sensor adjustment parameter may comprise calculating a value representative of the pixel intensity values ​​of at least a portion of an image of the succession of images, in particular a representative value resulting from the application of a statistical function to said pixel intensity values, determining a deviation between said representative value and a target value, comparing said deviation to a tolerance threshold and, depending on the result of the comparison, applying a new value of the sensor adjustment parameter.

[0028] The new value of the sensor adjustment parameter can be selected, at least as a function of said deviation and the temperature of the detection system, from a predetermined table of correspondence between a variation of the sensor adjustment parameter and a variation of the average intensity value of said part of the image.

[0029] The new value of the sensor adjustment parameter selected in the said temperature-dependent correspondence table allows precise adjustment of the sensor according to the implementation conditions of the detection system.

[0030] The target value may correspond to a voltage value between a minimum analog-to-digital conversion voltage value of the detection system and a maximum analog-to-digital conversion voltage value of the detection system.

[0031] The target value can be the average of the minimum analog-to-digital conversion voltage value of the detection system and the maximum analog-to-digital conversion voltage value of the detection system.

[0032] The target value can be determined based on calibration, particularly at the factory, and an estimated operational need.

[0033] The target value may correspond to the average of the minimum analog-to-digital conversion voltage value of the detection system and the maximum analog-to-digital conversion voltage value of the detection system corrected according to the estimated operational requirement.

[0034] The estimated operational need corresponds to a need to measure objects whose temperature is far from the average temperature of the scene. When the estimated operational measurement need corresponds to a need to measure objects whose temperature is higher than the average temperature of the scene, the target value is increased. When the estimated operational measurement need corresponds to a need to measure objects whose temperature is lower than the average temperature of the scene, the target value is decreased.

[0035] In a preferred embodiment, the target value is a statistical value calculated as a gray level of an unsaturated image provided by the detection system, in particular a statistical value resulting from the application of a statistical function on the gray level of the pixels of an unsaturated image provided by the detection system, for example the average of the gray level of an unsaturated image provided by the detection system.

[0036] Thus, the sensor adjustment parameter is modified by controlling the average intensity value of at least part of an image in the succession of images to a digital target value.

[0037] Modifying the sensor reference offset table may include determining and applying a new reference offset table iteratively determined by adding to a current sensor reference offset table a weighting of at least one predetermined sensor reference offset correction table, the weighting factor being determined according to the following law:

[0038] Where "i is the weighting factor calculated at a given calculation time (i) associated with a current image of the succession of images

[0039] ( | ) is the canonical scalar product;

[0040] HF() is a predetermined high-pass filter;

[0041] TOi is the predetermined sensor reference offset correction table implemented at calculation time (i); and li is the current image at calculation time (i). The weighting factor a ttherefore corresponds to a normalized correlation between a high-frequency signal of the current image and the predetermined sensor reference offset correction table implemented at the calculation instant (i). The new reference offset table determined iteratively therefore depends on the signal present in the images of the succession of images. The new reference offset table determined iteratively therefore depends on the modification of the sensor adjustment parameter, as well as on the modification of the sensor reference offset table applied to the images.

[0042] The predetermined sensor reference offset correction tables can be previously determined during calibrations of the detection system, particularly in the factory.

[0043] Predetermined sensor reference offset correction tables can be applied depending on the temperature of the sensing system.

[0044] Advantageously, the weighting factor is determined for the current image at the calculation time, the new reference offset table being applied for the image of the succession of images directly successive to the current image at the calculation time.

[0045] Compensating for variation in the average intensity of the images may include applying to at least a portion of an image of the succession of images provided by the detection system the opposite of a predetermined value of variation in average intensity between two images of a reference scene, the two images of the reference scene being provided by the detection system before and after an equivalent modification of the sensor adjustment parameter.

[0046] The equivalent modification of the sensor adjustment parameter is, for example, a modification of the same value of said adjustment parameter when the detection system is facing the reference scene, in particular when the detection system is implemented under identical measurement conditions.

[0047] Said predetermined value of average intensity variation can be selected, at least as a function of the temperature of the detection system and the modification of the adjustment parameter of the sensor, in a predetermined compensation table.

[0048] Said predetermined value of average intensity variation selected at least as a function of the temperature of the detection system and the modification of the sensor adjustment parameter in the predetermined compensation table allows precise adjustment of the sensor as a function of the conditions of implementation of the detection system.

[0049] Said predetermined value can also be selected according to the average intensity value of said part of the image in the predetermined compensation table.

[0050] Compensation for the average intensity variation of images can be applied to several successive images so that the compensation decreases from one image to the next.

[0051] The compensation value of the average intensity variation applied to successive images is, for example, calculated by applying a strictly decreasing function to said predetermined average intensity variation value to calculate the compensation values ​​to be applied to successive images.

[0052] Thus, the average intensity of the images provided by the detection system does not vary abruptly during the process.

[0053] The present invention also relates to a device for acquiring and processing images of a scene provided by a detection system comprising a sensor provided with a matrix of detectors sensitive to electromagnetic radiation coming from the scene, called "active", a plurality of detectors insensitive to said radiation, called "blind", as well as a reading circuit providing a response signal for each sensitive detector, the detection system providing a succession of images of the scene from the response signals, the device comprising: a means for modifying a sensor adjustment parameter;

[0054] - a means for modifying a sensor reference offset table applied to the images; and a means for compensating for a variation in the average intensity of the images caused by the modification of the sensor adjustment parameter and the sensor reference offset table applied to the images.

[0055] The device may comprise means for calculating an average intensity value of at least part of an image of the succession of images, means for determining a deviation between said average intensity value and a target value and means for comparing the deviation to a tolerance threshold.

[0056] The device may comprise a means for storing the predetermined table of correspondence between the variation of the sensor adjustment parameter and the variation of the average intensity value of said part of the image.

[0057] The device may include means for determining the new reference offset table.

[0058] The means for compensating for the variation in the average intensity of the images can be configured to apply to at least part of an image of the succession of images provided by the detection system the opposite of the predetermined value of variation in average intensity between two images of a reference scene.

[0059] The device may include means for storing a predetermined compensation table.

[0060] The means for compensating for variation in the average intensity of the images can be configured to apply the compensation to several successive images so that the compensation decreases from one image to the next.

[0061] The device may include a means for selecting an estimated operational need.

[0062] Brief description of the drawings

[0063] Other aims, characteristics and advantages of the invention will appear on reading the following description, given solely by way of non-limiting example and made with reference to the appended drawings in which: [Fig 1] schematically illustrates a detection system comprising a device according to the invention;

[0064] [Fig 2] schematically illustrates a method of acquiring and processing images according to the invention; and

[0065] [Fig 3] schematically illustrates the preliminary determination of tables according to the process.

[0066] Detailed description

[0067] Figure 1 schematically represents a detection system 2 observing a scene 4, for example an uncooled infrared detection system, also called an IRNR detector.

[0068] The detection system 2 can be a surveillance and / or monitoring system.

[0069] The detection system 2 comprises an optical sensor 6 provided with a matrix of detectors 8 sensitive to electromagnetic radiation coming from the scene 4, called “active” detectors 8, for example a matrix of bolometric detectors. The sensor 6 also comprises a plurality of detectors 10 insensitive to electromagnetic radiation coming from the scene 4, called “blind” detectors 10, and a reading circuit 12 providing a response signal for each sensitive detector 8 of the sensor 6.

[0070] The detection system 2 provides, from the response signals of the active detectors 8, a succession of images of the scene 4 forming a video stream of the scene 4.

[0071] Electromagnetic radiation includes infrared, even visible, terahertz or ultraviolet radiation.

[0072] The detection system 2 comprises an arrangement 14 of optical components making it possible to project the image of the scene 4 onto the matrix of active detectors 8 located in the focal plane of the arrangement of optical components 14. Such a detection system 2 is known as a “focal-plane array” (FPA) in English terms.

[0073] Each active detector 8 comprises a thermal resistance sensitive to the electromagnetic radiation of the scene 4. A current passing through the thermal resistance of each active detector 8 is representative of heating of said active detector 8 linked to the electromagnetic radiation of the scene 4.

[0074] Each blind detector 10 is a detector insensitive, or almost insensitive, to the radiation from the scene 4, generating a current, called compensation, representative of the non-useful current notably linked to the heating of the elements of the detection system 2.

[0075] The blind detectors 10 are arranged so that several blind detectors 10 are associated with each row or each column of the active detector matrix 8.

[0076] The difference between the current flowing through the thermal resistance of each active detector 8 and the compensation current produced by the blind detectors 10 is calculated in order to extract a useful analog signal from the scene 4.

[0077] This current difference is then integrated, in particular by filling a capacitor, over a period called “integration time”, thus generating a voltage constituting a useful integrated analog signal associated with each active detector 8.

[0078] The reading circuit 12 of the sensor 6 makes it possible to multiplex and sequentially supply the useful integrated analog signals associated with each active detector 8 of the matrix.

[0079] The reading circuit 12 comprises a trans-impedance capacitive amplifier making it possible to apply a trans-impedance capacitive amplification gain of the sensor 6, called “CTIA gain”, to the integrated useful analog signals.

[0080] The detection system 2 comprises implementation electronics (not shown), called “proximity electronics”, provided with at least one analog-digital converter configured to digitally convert the integrated useful analog signals to which the CTIA gain is applied in order to form the images of the scene 4, the images therefore being provided from the signals of the active detectors 8.

[0081] The proximity electronics is capable of polarizing the active detectors 8 by supplying a polarization voltage to the active detectors 8. The proximity electronics is capable of polarizing the blind detectors 10 by supplying a polarization voltage to the blind detectors 10, called “base voltage”.

[0082] The detection system 2 comprises an image acquisition and processing device 16 comprising a temperature sensor 18 capable of measuring the temperature of the detection system 2 or of the sensor 6 of the detection system 2. For example, the temperature sensor 18 is integrated into the sensor 6 of the detection system 2.

[0083] The image acquisition and processing device 16 comprises a computer 20 provided with a memory.

[0084] The image acquisition and processing device 16 also comprises a means 22 for modifying an adjustment parameter of the sensor 6, a means 24 for modifying a reference offset table of the sensor 6 applied to the images, a means 26 for compensating for a variation in the average intensity of the images provided by the detection system 2, a means 28 for calculating an average intensity value of at least one part of an image of the succession of images provided by the detection system 2, a means 30 for determining a difference between the average intensity value and a target value, a means 32 for comparing the difference to a tolerance threshold, a means 34 for storing a predetermined table of correspondence between a variation in the adjustment parameter of the sensor 6 and a variation in the average intensity value of the part of the image, a means 36 for determining a new reference offset table,a means 38 for storing a predetermined compensation table and a means 40 for selecting an estimated operational requirement.,

[0085] For example, the computer 20 of the image acquisition and processing device 16 comprises modules in the form of calculation codes, said means 22, 24, 26, 28, 30, 32, 34, 36, 38, 40 respectively comprising a module for modifying an adjustment parameter of the sensor 6, a module for modifying a reference offset table of the sensor 6 applied to the images, a module for compensating for a variation in the average intensity of the images provided by the detection system 2, a module for calculating an average intensity value of at least a part of an image of the succession of images provided by the detection system 2, a module for determining a difference between the average intensity value and a target value, a module for comparing the difference to a tolerance threshold,a module for storing a predetermined table of correspondence between a variation of the adjustment parameter of the sensor 6 and a variation of the average intensity value of the part of the image, a module for determining a new reference offset table, a module for storing a predetermined compensation table and a module for selecting an estimated operational requirement.,

[0086] The computer 20 also includes program code instructions for implementing an image acquisition and processing method.

[0087] Figure 2 schematically represents the image acquisition and processing process.

[0088] Optionally, the method starts with a step 42 of initializing the detection system 2 during which the temperature of the detection system 2 is measured to select and apply initial adjustment parameters of the sensor 6 and an initial reference offset table of the sensor 6.

[0089] For example, the memory of the computer 20 comprises adjustment parameters of the sensor 6 and reference offset tables predetermined, in particular in the factory, for several operating temperatures of the detection system 2. The temperature sensor 18 measures the temperature of the detection system 2 or of the sensor 6, then the modification means 22 selects from the memory of the computer 20 initial adjustment parameters of the sensor 6 adapted to the measured temperature, then applies said initial parameters to the sensor 6, the modification means 24 of the reference offset table of the sensor 6 selecting from the memory of the computer 20 an initial reference offset table of the sensor 6 adapted to the measured temperature, then applies said initial table to the sensor 6.The adjustment parameters of the sensor 6 include, for example, the bias voltage applied to the active detectors 8, the basing voltage applied to the blind detectors 10 and the CTIA gain applied by the reading circuit 12.

[0090] For example, the average temperature of the observed scene 4 is estimated to be close to the temperature of the detection system 2 during its initialization.

[0091] The initialization step 42 is, for example, carried out when the detection system 2 is powered up or regularly so that the detection system 2 also adapts to the heating of the components of the detection system 2 and to the variations in the scene 4 observed.

[0092] Advantageously, the initial parameters of the sensor 6 and the initial reference offset table of the sensor 6 are also selected according to the estimated operational requirement. The operational requirement can be estimated manually by an operator using the means 40 for selecting an operational requirement or fixed by design of the detection system 2.

[0093] The initial value of the CTIA gain adapted to the measured temperature and the estimated operational need makes it possible to reduce digital conversion noise.

[0094] The initial value of the bias voltage of the active detectors 8 adapted to the measured temperature and the estimated operational need makes it possible to reduce a measurement noise of the detection system 2, in particular a temporal noise of the sensor 6 corresponding to the variations of the response signals of the active detectors 8 when the detection system 2 observes a fixed scene. This measurement noise corresponds to the NETD, for “noise equivalent temperature difference”, in Anglo-Saxon terms.

[0095] The initial value of the base voltage adapted to the temperature and the estimated operational need reduces the risks of saturation of the objects of the scene 4, in particular of the objects corresponding to the estimated operational need. The initial reference offset table of the sensor 6 adapted to the measured temperature and the estimated operational need is, preferably, also adapted to the initial value of the bias voltage of the active detectors 8, to the initial value of the base voltage and to the initial value of the CTIA gain.

[0096] A step 44 of acquiring images of the scene 4 is then carried out during which the detection system 2 provides, from the response signals of the active detectors 8, a succession of images of the scene 4.

[0097] Then, a step 46 of processing the images of scene 4 is carried out, preferably simultaneously with the acquisition step 44.

[0098] Processing step 46 comprises a sub-step 48 of modifying at least one of the adjustment parameters of the sensor 6.

[0099] For example, during sub-step 48 the bias voltage applied to the active detectors 8 or the basing voltage or the value of the CTIA gain is modified.

[0100] For example, at a given calculation instant, one begins by calculating an average intensity value of the current image or of a part of the current image of the succession of images of the scene 4 provided by the detection system 2. For example, the calculation means 28 performs the calculation of the average value. Alternatively, one begins by calculating a value representative of the intensity values ​​of the pixels of said current image or part of the current image, in particular a representative value resulting from the application of a statistical function to said intensity values ​​of the pixels. For example, the calculation means 28 can be configured to calculate said representative value.

[0101] Then, a deviation between said average value and a target value is determined. For example, the determining means 30 calculates the deviation. Alternatively, the deviation is calculated between said representative value and the target value by the determining means 30.

[0102] The deviation is then compared to a tolerance threshold. For example, the comparison means 32 performs the comparison. If the deviation is greater than the tolerance threshold, the base voltage is modified to reduce the deviation. For example, the modification means 22 performs the modification of the base voltage.

[0103] The target value corresponds, for example, to a voltage value between a minimum analog-to-digital conversion voltage value of the detection system 2 and a maximum analog-to-digital conversion voltage value of the detection system 2.

[0104] Advantageously, the minimum and maximum analog-digital conversion voltage values ​​of the detection system 2 delimit a linear analog-digital conversion range of the detection system 2.

[0105] More precisely, the target value corresponds, for example, to the average of the minimum analog-to-digital conversion voltage value of system 2 and the maximum analog-to-digital conversion voltage value of system 2.

[0106] For example, the measured temperature of detection system 2 is +20°C. A minimum temperature to be detected in the observed scene 4 is estimated to be 0°C and a maximum temperature to be detected in the observed scene 4 is estimated to be +40°C, an object at the minimum temperature or at the maximum temperature to be detected should not saturate the image of scene 4 provided by detection system 2.The current bias voltage value of the active detectors 8 and the current CTIA gain value, applied to the sensor 6 when the detection system 2 provides the current image, make it possible to encompass, or at least to match, the temperature dynamics of the scene 4, namely a temperature variation between the minimum temperature value to be detected and the maximum temperature value to be detected, to the dynamics of the sensor 6, namely the voltage interval taken between the minimum analog-digital conversion voltage value of the system 2 and the maximum analog-digital conversion voltage value of the system 2.

[0107] For example, the minimum analog-to-digital conversion voltage value of system 2 is OV, the maximum analog-to-digital conversion voltage value of system 2 is 2V, and the average temperature of the observed scene 4 is +20°C. If the average intensity value calculated for the current image of scene 4 corresponds to a signal voltage of 2V, objects above +20°C in scene 4 may be saturated in the current image of scene 4. If the deviation exceeds the tolerance threshold, the basing voltage is modified so that the average intensity calculated for the next images of scene 4 approaches the target value, thus reducing the risk of saturation in the current image of objects above +20°C in the observed scene 4.

[0108] Advantageously, the target value is corrected according to the estimated operational need.

[0109] For example, it is estimated that hotter objects, for example at +60°C, need to be measured. The target value is then modified so that the target value corresponds to a value different from the average of the minimum analog-to-digital conversion voltage value of system 2 and the maximum analog-to-digital conversion voltage value of system 2 to reduce the risk of saturation of hotter objects.

[0110] Advantageously, the modification means 22 communicates with the storage means 34 to select a variation value of the adjustment parameter of the sensor 6, here a variation value of the basing voltage, as a function of the temperature of the detection system 2 and the difference in the predetermined correspondence table between the variation of the adjustment parameter of the sensor 6 and the variation of the average intensity value of the image. Then the modification means 22 applies said adjustment parameter of the sensor 6, here said variation value of the basing voltage.

[0111] Preferably, the target value is a numerical value, in particular a statistical value calculated as a gray level of an unsaturated image provided by the detection system.

[0112] For example, conversion coefficients between the variation in the basing voltage and the average variation in intensity of the image provided by the detection system 2 are recorded in the memory of the computer 20 for several temperatures of the detection system 2 during a prior calibration, in particular in the factory, of the detection system 2.

[0113] Advantageously, the tolerance threshold is determined as a function of a digital step of the adjustment parameter of the sensor 6. For example, the modification of the base voltage is applied only if the variation value of the base voltage to be applied is greater than the digital step of the base voltage values ​​recorded in the memory of the computer 20.

[0114] The processing step 46 also comprises a sub-step 50 of modifying the reference offset table of the sensor 6 applied to the images of the succession of images provided by the detection system 2 so that the new reference offset table of the sensor 6 is adapted to the new value of the adjustment parameter of the modified sensor 6.

[0115] For example, depending on the modification of the adjustment parameter of the sensor 6, the modification means 24 makes a selection in the memory of the computer 20 of a reference offset table of the sensor 6 adapted to the new value of the adjustment parameter of the modified sensor 6.

[0116] In a preferred variant, the determination means 36 is capable of iteratively determining the new reference offset table of the sensor 6. For example, the determination means 36 performs the calculation of a weighting factor at a given calculation time from the current image according to the following law:

[0117] Where ( | ) is the canonical dot product;

[0118] HF() is a predetermined high-pass filter;

[0119] TOi is a predetermined reference offset correction table of the sensor 6 implemented at the calculation time (i); and li is the current image at the calculation time (i). The determining means 36 then performs the sum of the current reference offset table of the sensor 6, applied to the sensor 6 when the detection system 2 provides the current image, and the predetermined reference offset correction table of the sensor 6 implemented at the calculation time, coefficiented by the weighting factor calculated at the calculation time in order to obtain the new reference offset table of sensor 6.

[0120] Then, the modification means 24 communicates with the determination means 36 and applies the new reference offset table of the sensor 6 to the image directly following the current image.

[0121] Advantageously, the memory of the calculator 20 comprises several predetermined tables for correcting the reference offset of the sensor 6, said correction tables being previously determined for several temperatures of the detection system 2. In this case, the determination means 36 performs calculations of weighting factors a on several successive images. L for several predetermined tables of reference offset correction of the sensor 6 adapted to the temperature of the detection system 2.

[0122] In another variant, the modification means 24 is capable of modifying the reference offset table of the sensor 6 applied to the images by a method of single-point calibration and / or wavelet row processing and / or wavelet column processing and / or Weiss calibration and / or calibration based on optical flow.

[0123] Finally, the processing step 46 comprises a sub-step 52 of compensating for the variation in the average intensity of the images generated by the modification of the sensor adjustment parameter 6 and of the reference offset table of the sensor 6 applied to the images of the succession of images provided by the detection system 2.

[0124] Advantageously, simultaneously with the application of the new reference shift table to the image directly following the current image, the opposite of a predetermined value of average intensity variation between two images of a reference scene is applied to said directly following image, the two images of the reference scene being provided by the detection system 2 before and after an equivalent modification of the adjustment parameter of the sensor 6. The reference scene is, for example, a scene of known constant temperature. The equivalent modification is, for example, an equal modification of the adjustment parameter of the sensor 6 or a modification representative of the modification of the adjustment parameter of the sensor 6.

[0125] For example, the compensation means 26 communicates with the storage means 38 to carry out a selection of said predetermined value in the predetermined compensation table as a function of the temperature of the detection system 2 and the modification of the adjustment parameter of the sensor 6. Then the compensation means 26 carries out the application of said predetermined value to said directly successive image.

[0126] Advantageously, the compensation means 26 applies said predetermined value to several successive images in a decreasing manner, for example with a linear or exponential decrease, so that the variation in intensity from one image to another of said successive images is gentle for the operator observing said successive images of the scene provided by the detection system 2. In particular, so that the difference in average intensity level between two successive images is less than a gray level value defined for the comfort of the operator.

[0127] Figure 3 schematically represents the steps of preliminary calibration of the detection system 2, particularly in the factory.

[0128] First of all, a step 54 is carried out for recording reference offset tables of the sensor 6 in the memory of the computer 20.

[0129] For example, for several bias voltage values ​​of the active detectors 8, for several basing voltage values, for several CTIA gain values ​​and for several temperature values ​​of the detection system 2 or of the sensor 6, reference offset tables of the sensor 6 are established and recorded in order to compensate for the spatial non-uniformities of the images provided by the detection system 2.

[0130] A step 56 is then carried out for recording a table of CTIA gain values ​​in the memory of the computer 20.

[0131] For example, for several temperatures of the detection system 2 or of the sensor 6, and for several reference scenes, CTIA gain values ​​selected to reduce the digital conversion noise of the detection system 2 are recorded.

[0132] A step 58 is also carried out for recording a table of polarization voltage values ​​of the active detectors 8 in the memory of the computer 20.

[0133] For example, for several temperatures of the detection system 2 or of the sensor 6, and for several reference scenes, bias voltage values ​​of the active detectors 8 are recorded, selected to reduce the measurement noise of the detection system 2, in other words to reduce the NETD.

[0134] Preferably, the CTIA gain and bias voltage values ​​of the active detectors 8 are selected and recorded simultaneously so as to jointly reduce the measurement and digital conversion noise values ​​of the detection system 2.

[0135] A step 60 of recording a table of base voltage values ​​in the memory of the computer 20 is also carried out.

[0136] For example, for several temperatures of the detection system 2 or of the sensor 6, for several reference scenes, for several bias voltages of the active detectors 8 and for several CTIA gains, selected basing voltage values ​​are recorded to adapt the temperature dynamics of the measured scene 4 to the dynamics of the sensor 6.

[0137] A step 62 is also carried out for recording the correspondence table between the variation of the adjustment parameter of the sensor 6 and the variation of the average intensity value of the image in the storage means 34. For example, the correspondence table between the variation of the basing voltage and the variation of the average intensity value of the image is recorded in the storage means 34.

[0138] For example, for several temperatures of the detection system 2 or of the sensor 6, for several reference scenes, for several bias voltage values ​​of the active detectors 8, for several basing voltage values, for several CTIA gains and for several reference offset tables of the sensor 6, digital values ​​of variation of the adjustment parameter are recorded, for example digital values ​​of variation of the basing voltage, making it possible to obtain a certain variation of the average intensity value of the image or of a part of the image.

[0139] Thus, when the deviation is determined during the sub-step 48 of modifying the adjustment parameter of the sensor 6, the modification means 22 communicates with the storage means 34 to carry out the modification of the adjustment parameter adapted to the deviation.

[0140] A step 64 of recording reference offset correction tables of the sensor 6 in the memory of the computer 20 is also carried out.

[0141] For example, the detection system 2 is placed opposite a reference scene, in particular a black body of known temperature. Then the temperature of the detection system 2 is controlled at a given temperature. The bias voltage of the active detectors 8, the basing voltage, the value of the CTIA gain and the value of the reference offset table of the sensor 6 are adjusted. The average intensity value of the image or of a part of the image of the reference scene is measured. Then the adjustment parameter of the sensor 6 is varied and the reference offset correction tables of the sensor 6 to be applied are determined to correct the reference offset defects of the sensor 6 occurring following the variation of the adjustment parameter of the sensor 6.

[0142] Finally, a step 66 of recording the compensation table in the storage means 38 is carried out. For example, for several temperatures of the detection system 2 or of the sensor 6, for several reference scenes, for several bias voltage values ​​of the active detectors 8, for several basing voltage values, for several CTIA gains and for several reference offset tables of the sensor 6, variation values ​​of the average intensity value of the image or of a part of the image corresponding to a digital variation of the adjustment parameter, for example corresponding to the digital adjustment step of the adjustment parameter, for example corresponding to a digital variation of the basing voltage, are recorded.

[0143] Thus, during the compensation sub-step 52, the compensation means 32 communicates with the storage means 38 to perform compensation for the average intensity variation of the images adapted to the modification of the adjustment parameter of the sensor 6 carried out during the sub-step 48 of modification of the adjustment parameter of the sensor 6 and adapted to the modification of the reference offset table of the sensor 6 carried out during the sub-step 50 of modification of the reference offset table of the sensor 6.

[0144] Advantageously, said tables are established in the form of functions or are established for temperature segments, for example for temperatures between -40°C and -10°C, between -10°C and +40°C and between +40°C and +70°C, in particular in order to reduce the quantity of recorded data.

Claims

CLAIMS 1. Method for acquiring and processing images of a scene (4) provided by a detection system (2) comprising a sensor (6) provided with a matrix of detectors sensitive (8) to electromagnetic radiation coming from the scene (4), called "active", a plurality of detectors insensitive (10) to said radiation, called "blind", as well as a reading circuit (12) providing a response signal for each sensitive detector (8), the detection system (2) providing a succession of images of the scene (4) from the response signals, characterized in that it comprises: the modification of an adjustment parameter of the sensor (6); the modification of a reference offset table of the sensor (6) applied to the images; and the compensation of a variation in the average intensity of the images generated by the modification of the adjustment parameter of the sensor (6) and of the reference offset table of the sensor (6) applied to the images.

2. Method according to claim 1, in which the adjustment parameter of the sensor (6) comprises a bias voltage applied to the insensitive detectors (10) called “base voltage” and / or a value of a trans-impedance capacitive amplification gain of the sensor (6) called “CTI A gain”.

3. Method according to one of claims 1 and 2, in which the modification of the adjustment parameter of the sensor (6) comprises the calculation of a value representative of the intensity values ​​of the pixels of at least a part of an image of the succession of images, the determination of a deviation between said representative value and a target value, the comparison of said deviation with a tolerance threshold and, according to the result of the comparison, the application of a new value of the adjustment parameter of the sensor (6).

4. Method according to claim 3, in which the new value of the adjustment parameter of the sensor (6) is selected, at least as a function of said deviation and the temperature of the detection system (2), in a predetermined table of correspondence between a variation in the adjustment parameter of the sensor (6) and a variation in the average intensity value of said part of the image.

5. Method according to one of claims 3 and 4, in which the target value is determined based on a factory calibration and an estimated operational need.

6. Method according to any one of claims 1 to 5, wherein the modification of the reference offset table of the sensor (6) comprises the determination and application of a new reference offset table determined iteratively by adding to a current reference offset table of the sensor (6) a weighting of at least one predetermined reference offset correction table of the sensor (6), the weighting factor being determined according to the following law: Where "i is the weighting factor calculated at a given calculation time (i) associated with a current image of the succession of images ( | ) is the canonical scalar product; HF() is a predetermined high-pass filter; TOi is the predetermined sensor reference offset correction table (6) implemented at the calculation time (i); and li is the current image at the calculation time (i).

7. Method according to any one of claims 1 to 6, in which the compensation for the variation in the average intensity of the images comprises the application to at least part of an image of the succession of images provided by the detection system (2) of the opposite of a predetermined value of variation in average intensity between two images of a reference scene, the two images of the reference scene being provided by the detection system (2) before and after an equivalent modification of the adjustment parameter of the sensor (6).

8. Method according to claim 7, wherein said predetermined value of average intensity variation is selected, at least as a function of the temperature of the detection system (2) and the modification of the adjustment parameter of the sensor (6), in a predetermined compensation table.

9. Method according to any one of claims 1 to 8, in which the compensation for the variation in average intensity of the images is applied to several successive images so that the compensation decreases from one image to the next.

10. Device (16) for acquiring and processing images of a scene (4) provided by a detection system (2) comprising a sensor (6) provided with a matrix of detectors sensitive (8) to electromagnetic radiation coming from the scene (4), called "active", a plurality of detectors insensitive (10) to said radiation, called "blind", as well as a reading circuit (12) providing a response signal for each sensitive detector (8), the detection system (2) providing a succession of images of the scene (4) from the response signals, characterized in that it comprises: a means for modifying an adjustment parameter of the sensor (6); a means for modifying a reference offset table of the sensor (6) applied to the images;and means for compensating for a variation in the average intensity of the images caused by the modification of the adjustment parameter of the sensor (6) and of the reference offset table of the sensor (6) applied to the images.;

Images