Method of processing raw images collected by bolometer detectors and associated equipment

Abstract

A method for processing a raw image, characterized in that it is performed by a computer based on reference measurements Pix REF (i,j) and second reference measurements Pix REF (i,j) associated respectively with the first bolometer Bol1(i,j) and the second bolometer Bol2(i,j), the method comprising: a) a correlation step of correlating the first raw measurements Pix1(i,j) and the first reference measurements Pix 1REF (i,j), b) a step of correcting the raw image comprising calculating, for each bolometer Bol(i,j), a corrected measurement Pix 2REF (i,j) of a corrected image of the raw image based on the reference measurements Pix 1REF (i,j) and the result of step (a). REF (i,j) of a corrected image of the raw image based on the reference measurements Pix Cor (i,j) and the result of step (a).

Description

Technical Field

[0001] This invention relates to a method for processing raw images collected by a detector equipped with a matrix array of calorimeters. The processing method according to the invention is particularly aimed at correcting inhomogeneities caused by discrete characteristics in the calorimeters of the detector.

[0002] The present invention also relates to a computer program capable of implementing all the steps of the method according to the present invention.

[0003] Finally, the present invention relates to a detector including a calorimeter and a computer implementing the computer program. Background Technology

[0004] Infrared detectors (or imagers) known from existing technology typically consist of a calorimeter organized in the form of a matrix array of n rows and m columns.

[0005] When exposed to a scene to acquire images, these calorimeters, which are sensitive to scene temperature, experience changes in resistance. In other words, the current flowing in each calorimeter depends on the scene temperature, but also on the ambient temperature.

[0006] In particular, the measured value Sp(i,j) of the radiometer Bol(i,j) of the radiometer matrix array develops according to the following rules:

[0007]

[0008] in:

[0009] -T amb It is related to the ambient temperature, and also to the temperature of the imager;

[0010] -T scene The temperature of the scene as seen by radiometers i and j;

[0011] -R esp It refers to the response capability of the calorimeter;

[0012] -S 0，Tamb It is the output value of the calorimeter when the scene temperature equals the imager temperature.

[0013] Item R esp This reflects the ability of a calorimeter to convert an input signal (especially radiation) into an electrical signal, particularly an electric current. Item R esp The resistivity of the calorimeter is closely related to the materials and architecture used in its design. This resistivity varies as a function of the ambient temperature and the difference between that ambient temperature and the scene temperature. Calculating the scene temperature requires knowing the imager's Te. amb The detector is usually equipped with a temperature sensor to measure the temperature.

[0014] In practice, these detectors are also equipped with additional calorimeters, called blind calorimeters, which are not exposed to the environment. Their resistivity, and therefore the current flowing through them, depends only on the ambient temperature T. amb .

[0015] Therefore, differential measurements of the current flowing through blind calorimeters and calorimeter matrix arrays can be used to infer changes in resistivity.

[0016] Typically, each column of the calorimeter matrix array is associated with a blind calorimeter implemented for each of the calorimeters in that column during differential measurements. However, other configurations are conceivable, particularly for combining a single blind calorimeter into multiple columns of calorimeters.

[0017] The original image of the scene can be obtained using this device. Figure 1 It is usually unusable and requires additional processing.

[0018] In particular, Figure 1 The image shown illustrates the positioning of the detector's calorimeters, and more specifically, the non-uniformity ("pixelation effect"). This effect stems from the significant dispersion of resistance from one calorimeter to another.

[0019] The image also has a columnar appearance, which is caused by the resistance dispersion between blind calorimeters.

[0020] Various solutions have been considered to overcome these problems.

[0021] It has been specifically proposed to implement a mechanical shutter on the detector. The mechanical shutter is placed in front of the detector to collect a reference image relative to the ambient temperature, which is then subtracted from the scene image.

[0022] This layout, which is relatively simple in principle, is not entirely satisfactory.

[0023] Specifically, implementing a shutter and its associated mechanical devices presents problems in terms of cost and size.

[0024] In addition, the reference image must be refreshed if the ambient temperature changes.

[0025] Alternatively, characterization of the response at a reference temperature has been proposed, particularly the temperature of the detector, and especially the temperature of each of its calorimeters.

[0026] This characterization includes reference measurements at various temperatures, where all the calorimeters of the detector are closed with the shutter on.

[0027] Then, reference measurements can determine the temperature variation of each calorimeter, thereby constructing a calibration table stored in the detector's storage space.

[0028] Therefore, during operation, the detector corrects the original image by subtracting the value obtained through interpolation based on the calibration table for each calorimeter.

[0029] However, this solution, which can reduce the effects of non-uniformity in the image from one calorimeter to another, is not satisfactory.

[0030] Specifically, the process of obtaining reference measurements is lengthy and incurs additional production costs for the detector.

[0031] Furthermore, due to the associated costs, it is undesirable to dedicate storage space to storing calibration tables.

[0032] A third algorithm-based method for correcting image non-uniformity is proposed in documents [1] and [2] cited at the end of the specification.

[0033] The methods known in the prior art are also unsatisfactory.

[0034] Specifically, these methods are often cumbersome to implement, and their robustness is questionable.

[0035] Furthermore, these methods require cumbersome computing devices, which further reduces the cost of implementing the detectors they employ.

[0036] Finally, a fourth method is proposed in document [3] cited at the end of the specification for correcting the raw measurements collected by the calorimeter by determining the reference matrix array. This method is relatively robust for correcting pixelation and columnar appearance that are easily observed in the raw images.

[0037] However, these two aspects are corrected through two separate procedures, so the fourth approach would benefit from simplification.

[0038] One object of the present invention is to provide a method for processing images collected by a detector with a calorimeter, which is easier to implement and does not require the attachment of mechanical parts and the associated robotics.

[0039] Another object of the present invention is to provide a method for correcting columnar effects observed in the original image. Summary of the Invention

[0040] The object of the present invention is achieved at least in part by the following method: a method for processing an original image, characterized in that the original measurement value Pix(i,j) includes a first original measurement value Pix1(i,j) and a second original measurement value Pix2(i,j), wherein the first original measurement value Pix1(i,j) and the second original measurement value Pix2(i,j) are respectively determined by a detector in n rows (L... i ) and m columns (C j The first calorimeter Bol1(i,j) and the second calorimeter Bol2(i,j) in the group of calorimeters Bol(i,j) arranged in the form of a matrix array are collected. During the acquisition of the raw measurement value Pix(i,j), the first calorimeter Bol1(i,j) is closed, and the data is collected by the computer based on the reference measurement value Pix. REF The method is executed using (i, j), with the reference measurement Pix as the reference. REF (i, j) includes the first reference measurement value Pix associated with the first calorimeter Bol1(i, j) and the second calorimeter Bol2(i, j), respectively. 1REF (i, j) and the second reference measurement Pix 2REF (i, j), the method includes:

[0041] a) Related steps: For each first calorimeter Bol1(i,j), the first original measurement value Pix1(i,j) and the first reference measurement value Pix1(i,j) are... 1REF (i, j) association;

[0042] b) The step of correcting the original image, which includes based on the reference measurement value Pix REF Based on the results of (i, j) and related steps (a), calculate the corrected measurement value Pix for at least each second calorimeter Bol(i, j) of the corrected image. Cor (i, j).

[0043] According to one implementation, the relevant step (a) includes linear regression, such that the first original measurement value Pix1(i,j) and the first reference measurement value Pix 1REF (i, j) satisfy the following relationship:

[0044] Pix1(i,j)=α.Pix 1REF (i, j) + Pix offset

[0045] Among them α and Pix offset It is a term determined during linear regression.

[0046] According to one embodiment, the calibration step (b) includes calculating the calibration measurement value Pix for at least each second calorimeter Bol(i,j) according to the following relationship. Cor(i, j):

[0047] Pix Cor (i,j)=Pix(i,j)-α.Pix REF (i,j)-Pix offset

[0048] According to one implementation, the reference measurement value Pix REF (i, j) are stored in the detector's storage space.

[0049] According to one embodiment, the detector is provided with a lens mounted on an aperture that closes the first calorimeter Bol1(i,j) at a corner of the detector.

[0050] According to one implementation, the reference measurement value Pix REF (i, j) was obtained during the detector calibration step by using a shutter to shield all calorimeters, with the shutter maintained at the same temperature as the detector.

[0051] According to one implementation, the reference measurement value Pix REF (i, j) is the average measurement of the signal actually collected by the calorimeter during the calibration step.

[0052] According to one embodiment, the detector also includes a blind calorimeter, Bol. blind (j), each blind calorimeter Bol blind (i,j) is used to perform differential measurements on at least one column of radiometers Bol(i,j) for its specific purpose; each blind radiometer Bol blind (i, j) is advantageously related to the single column (C) j The radiation calorimeter (Bol(i,j)) is associated with it.

[0053] The present invention also relates to a computer program that, when executed by a computer, causes the implementation of the method according to the invention.

[0054] The present invention also relates to an upper part, comprising:

[0055] - Detector, which is arranged in n rows (Li) and m columns (C) j The radiometers Bol(i,j) are arranged in a matrix array form. The radiometers include a first radiometer Bol1(i,j) and a second radiometer Bol2(i,j). The first radiometer Bol1(i,j) is enclosed.

[0056] - A computer, which is equipped with a computer program according to the present invention. Attached Figure Description

[0057] Other features and advantages will become apparent from the following description of the method for processing images by way of non-limiting example and with reference to the accompanying drawings, wherein:

[0058] Figure 1 The original image of the scene was obtained by a detector with a radiometer arranged in a matrix array of 80 rows by 80 columns;

[0059] Figure 2 This is a schematic diagram of a detector having a lens mounted on an aperture, which can be implemented according to the present invention;

[0060] Figure 3 Pix is ​​used as the first reference measurement. 1REF A plot-based depiction of the first raw measurement value Pix1(i,j) (on the vertical axis) of the function (i,j) (horizontal axis);

[0061] Figure 4a The original image is formed by the original measurement values ​​Pix(i,j) collected by the detector according to the present invention;

[0062] Figure 4b According to the present invention, the item is a correction measurement value Pix Cor The corrected image is formed by (i, j). Detailed Implementation

[0063] Now refer to Figures 1 to 4b This invention is described.

[0064] Figure 2 A detector 1 equipped with multiple calorimeters is shown, denoted as Bol(i,j), with n rows (denoted as "L"). i () and column m (represented as "C") j Arranged in the form of a matrix array of ").

[0065] The radiometer with index i,j corresponds to the radiometer located at the intersection of row i and column j.

[0066] In the calorimeter Bol(i,j), the first calorimeter Bol1(i,j) and the second calorimeter Bol2(i,j) can be distinguished.

[0067] Specifically, the calorimeters Bol(i,j) are arranged such that the first calorimeter Bol1(i,j) is turned off, while the second calorimeter Bol2(i,j) is exposed to the scene to be imaged.

[0068] In other words, only the second calorimeter Bol2(i,j) is sensitive to (thermal) radiation from the scene to be imaged, while the first calorimeter Bol1(i,j) does not receive or sense any radiation from the scene.

[0069] Therefore, according to a claim of the invention, the definition of the first calorimeter Bol1(i,j) may involve selecting the first calorimeter Bol1(i,j) from all calorimeters Bol1(i,j) of detector 1. In this respect, the selection may be based on considerations related to the effectiveness of turning off the first calorimeter Bol1(i,j).

[0070] It should also be understood that the closure of the first calorimeter Bol1(i,j) is permanent, and this can be achieved by a specific fixed element (e.g., an aperture described in the remainder of the specification) that forms part of the detector. "Fixed element" should be understood to mean that the element is immovable and non-motorizable.

[0071] The first calorimeter Bol1(i,j) can, for example, be enclosed by an aperture with a lens 2 mounted on it, which is positioned between the scene and the detector 1. In this configuration, the first calorimeter Bol1(i,j) is positioned in a corner of the detector. However, the invention is not limited to this layout, and the first calorimeter Bol1(i,j) can, for example, form a complete column of calorimeters, particularly a column adjacent to the calorimeter matrix array.

[0072] It should also be understood that the first calorimeter, Bol1(i,j), is only sensitive to radiation associated with the aperture that is turned off during data acquisition, which is advantageously at the same temperature as detector 1, because it is shielded.

[0073] Figure 2 The detectors in the system may also include multiple blind calorimeters (Bol). BLIND (j).

[0074] The detector 1 also includes a computer 4 with a computing processor for performing the various steps of the method according to the invention. The computer may also include storage space for storing raw measurements, reference measurements, or intermediate measurements.

[0075] Finally, detector 1 may include temperature detector 5, which is designed to evaluate a reference temperature, and in particular the temperature of the detector, denoted as T. amb Temperature detectors may include, for example, a PN junction.

[0076] According to the method of the present invention, defects in the raw images collected by detector 1 are corrected. The raw images are particularly characterized by the raw measurement set Pix(i,j), which represents the temperature T of the scene observed by each calorimeter i,j. scene These raw measurements, Pix(i,j), originate from the signal S actually transmitted by each calorimeter. p (i, j), and the observed scene is represented by the following formula:

[0077]

[0078] in:

[0079] -T amb It is the temperature of the imager;

[0080] -T scene The temperature of the scene as seen by radiometers i and j;

[0081] -R esp (i, j) represents the response capability of the calorimeter;

[0082] -S 0，Tamb It is the output value of the radiative calorimeter when the scene temperature equals the ambient temperature.

[0083] It should be understood that temperature T amb It is the temperature of each calorimeter of the imager, and in the case of thermal equilibrium with the surrounding environment where the imager is located, it is the same as the ambient temperature.

[0084] The response capability R of each calorimeter Bol(i,j) esp (i, j) can be determined during the production of the detector or before its first use. Specifically, determining the response capability Bol(i, j) can involve first and second calibration measurements. The first calibration measurement can particularly include exposing the detector, and more particularly all calorimeters, to an object or blackbody maintained at a first temperature advantageously equal to the temperature of the calorimeter, denoted below as T. amb The second calibration measurement may involve exposing the detector a second time to another blackbody maintained at a second temperature, different from the first temperature, for example, higher than the first temperature. Simply, knowing the reference temperature, and in particular the detector temperature, as well as the first and second temperatures, the response capability R of each calorimeter Bol(i,j) can be derived. esp (i, j). Response capability R esp (i, j) are specifically stored in the storage space of computer 4.

[0085] The calibration procedure remains easy to perform and does not require the implementation of algorithms that consume large amounts of memory and computational resources.

[0086] Figure 1 An example image formed from the original measurements Pix(i,j) is shown. This is discussed in the "Prior Art" section. Figure 1 It exhibits uniform defects and a columnar appearance.

[0087] The non-uniformity is mainly caused by the dispersion of resistance in the calorimeter that forms the detector.

[0088] Darker areas may also be observed in the corners of the image. These correspond to the original measurements Pix(i,j) and are called the first original measurements Pix1(i,j), collected by the first calorimeter Bol1(i,j). These dark areas are caused by the first calorimeter Bol1(i,j) being closed by an aperture with a lens attached, and will be used to correct for defects actually observed in the original image.

[0089] A brighter central region can also be seen corresponding to the original measurement value Pix(i,j) of the second calorimeter Bol2(i,j) (referred to as the second original measurement value Pix2(i,j)).

[0090] at last, Figure 1 It also exhibits a columnar appearance, which is due to the blind radiometric calorimeter Bol. BLIND (j) is caused by resistance dispersion. In this respect, each blind calorimeter Bol BLIND (j) Differential measurements are performed on at least one of its specific columns of radiometers; each blind radiometer Bol BLIND (j) Favorable with separate (C) j The radiation calorimeter Bol(i,j) is associated with this.

[0091] It should be noted that the column effect will not be observed if all columns are associated with the same blind calorimeter. However, this configuration is undesirable because any defect in this single blind calorimeter will render the entire detector inoperable.

[0092] The method according to the invention includes a first original measurement value Pix1(i,j) and a first reference measurement value Pix. 1REF The relevant steps (a) between (i, j).

[0093] First reference measurement Pix 1REF (i, j) Specifically, the measurements collected by each of the first calorimeters Bol1(i, j) through the application of a mask or shutter cover the entire detector 1 (and thus all the first and second calorimeters) and are kept at a reference temperature, for example, at the same temperature as said detector 1, such as 20°C.

[0094] The mask or shutter should not be confused with the aperture mentioned above. Specifically, a mask is a component placed in front of the detector (radiometer) for calibration purposes or to determine the technical characteristics of the radiometer during its production. This mask is not considered when using the detector.

[0095] It is understandable that, in obtaining the first reference measurement value Pix 1REF At (i, j), the second calorimeter Pix 2REF(i, j) also obtains the second reference measurement value Pix 2REF (i, j). These second reference measurements Pix 2REF (i, j) and the first reference measurement Pix 1REF Like (i, j), these correspond to images of a scene maintained at a reference temperature (e.g., the same as the detector temperature). These second reference measurements are Pix. 2REF (i, j) are used particularly during step (b) of the invention as described above.

[0096] First reference measurement Pix 1REF (i, j) and the second reference measurement Pix 2REF (i, j) can be stored in the storage space of computer 4. These reference measurements can be obtained during the calibration procedure, and advantageously simultaneously with the first calibration measurements described above.

[0097] According to the invention, step (a) can involve linear regression. Specifically, step (a) can involve determining the coefficients α and Pix via linear regression. offset In particular, based on the first reference measurement value Pix 1REF (i, j) are determined such that:

[0098] Pix1(i,j)=α.Pix 1REF (i, j) + Pix offset

[0099] Figure 3 The first reference measurement value, Pix, is shown. 1REF A graph of the first original measure Pix1(i,j) (vertical axis) of the function (i,j) (horizontal axis). A linear trend can be clearly observed, and therefore the choice of correlation can be demonstrated by linear regression.

[0100] This was unexpected, as no technical element could have predicted such behavior. Instead, given the numerous technical parameters defining the operation of a calorimeter, a more volatile trend was anticipated.

[0101] Especially in the absence of any prior correction or calibration of the columnar appearance of the first reference measurement Pix1(i,j), the correlation is obtained through linear regression, thus still yielding a satisfactory processing method that requires less computational resources.

[0102] The method according to the invention further includes step (b) of correcting the original measurement value Pix(i,j).

[0103] Step (b) of correcting the original image specifically includes correcting the reference measurement value Pix. REFBased on the results of (i, j) and related steps (a), calculate the corrected measurement value Pix for the corrected image of each calorimeter Bol(i, j). Cor (i, j).

[0104] Specifically, step (b) may involve calculating the corrected measurement value Pix. Cor (i, j). This correction of the original measurements may involve all the original measurements Pix(i, j) or only the second original measurement Pix2(i, j) representing the scene. It should be noted that the second alternative is still more advantageous in terms of computational resources.

[0105] Correction measurement value Pix Cor (i, j) can be calculated using the following relationship:

[0106] Pix Cor (i,j)=Pix(i,j)-α.Pix REF (i,j)-Pix offset

[0107] Therefore, calculate the corrected measurement value Pix. Cor (i, j) depends on the reference measurement Pix REF (i, j) without requiring any prior correction or calibration.

[0108] in this regard, Figure 4a and 4b The effect of the correction according to the present invention on the image collected by detector 1 from the scene is explained.

[0109] In particular, Figure 4a This is the original image of the scene, representing an individual, collected by detector 1. However, pixelation and columnar effects prevent the individual from being distinguished in this image.

[0110] In its regard, Figure 4b An image corrected according to the present invention is shown. The image no longer presents a pixelated and columnar appearance, and the individual can be clearly distinguished.

[0111] Therefore, the method according to the invention can correct inherent defects in images formed by a calorimeter detector. In particular, the method according to the invention can correct pixelation and columnar effects commonly observed in images formed by a calorimeter detector.

[0112] The method according to the present invention exhibits relatively low consumption of computing resources, while maintaining high robustness and efficiency.

[0113] Furthermore, this method does not involve any additional mechanical components for setting up automata, thereby reducing its manufacturing and usage costs.

[0114] The present invention also relates to a computer program that, when executed by a computer, causes the implementation of the correction method according to the present invention.

[0115] The present invention also relates to an apparatus comprising:

[0116] - Detector, which is configured with n rows (L) i ) and m columns (C j The radiometers Bol(i,j) are arranged in a matrix array. The radiometers include a first radiometer Bol1(i,j) and a second radiometer Bol2(i,j). The first radiometer Bol1(i,j) is enclosed.

[0117] - A computer equipped with a computer program according to the present invention.

[0118] References

[0119] [1]EP2940991B1.

[0120] [2]US2010237245A1.

[0121] [3]EP 3594643A1.

Claims

1. A method for processing raw images, characterized in that, The raw measurement value Pix(i,j) includes a first raw measurement value Pix1(i,j) and a second raw measurement value Pix2(i,j), which are respectively generated by the detector in n rows (L i ) and m columns (C j The first calorimeter Bol1(i,j) and the second calorimeter Bol2(i,j) in a group of calorimeters Bol(i,j) arranged in the form of a matrix array are collected. During the acquisition of the original measurement value Pix(i,j), the first calorimeter Bol1(i,j) is closed and the data is collected by the computer based on the reference measurement value Pix. REF The method is performed using (i,j), and the reference measurement value Pix is ​​used. REF (i,j) includes the first reference measurement value Pix associated with the first calorimeter Bol1(i,j) and the second calorimeter Bol2(i,j), respectively. 1REF (i,j) and the second reference measurement Pix 2REF (i,j), the reference measurement value Pix REF (i,j) are obtained during the detector calibration step by shielding all calorimeters with a shutter that is maintained at a reference temperature, the method comprising: a) Related steps, including the first original measurement value Pix1(i,j) and the first reference measurement value Pix 1REF The linear regression between (i,j) and each first calorimeter Bol1(i,j) satisfies the following relationship: Among them α and Pix offset These are terms determined during linear regression; b) The step of correcting the original image, which includes based on the reference measurement value Pix REF Based on the results of (i,j) and the related step a), calculate the corrected measurement value Pix for at least each second calorimeter Bol(i,j) of the corrected image. Cor (i,j), The calibration step b) includes calculating the calibration measurement value Pix for at least each second calorimeter Bol(i,j) according to the following relationship. Cor (i,j): 。 2. The method according to claim 1, wherein, The reference measurement value Pix REF (i,j) are stored in the detector's storage space.

3. The method according to claim 1, wherein, The detector is equipped with a lens mounted on an aperture, which closes the first calorimeter Bol1(i,j) at a corner of the detector.

4. The method according to claim 1, wherein, The reference temperature is equal to the detector temperature.

5. The method according to claim 4, wherein, The reference measurement value Pix REF (i,j) is the average measurement of the signal actually collected by the calorimeter during the calibration step.

6. The method according to claim 1, wherein, The detector also includes a blind calorimeter, Bol. blind (j), each blind calorimeter Bol blind (i,j) is used to perform differential measurements on at least one column of radiometers Bol(i,j) for its specific purpose; each blind radiometer Bol blind (i,j) and single column (C) j The radiation calorimeter (Bol(i,j)) is associated with it.

7. An apparatus comprising: - Detector, which is configured with n rows (L i ) and m columns (C j A radiometer Bol(i,j) is arranged in a matrix array, the radiometer including a first radiometer Bol1(i,j) and a second radiometer Bol2(i,j), the first radiometer Bol1(i,j) being enclosed. - A computer that performs the method according to claim 1.

Images