WO2010049569A1

WO2010049569A1 - Methods and systems for analysing sealing in fluid deposits

Info

Publication number: WO2010049569A1
Application number: PCT/ES2009/070473
Authority: WO
Inventors: Antonio J. DE CASTRO GONZÁLEZ; Fernando LÓPEZ MARTÍNEZ; Joaquín SÁNCHEZ TIRADO; Jose Manuel Aranda Gallego; Juan MELÉNDEZ SÁNCHEZ
Original assignee: Eads Construcciones Aeronauticas, S.A
Priority date: 2008-10-31
Filing date: 2009-10-29
Publication date: 2010-05-06
Also published as: ES2349911A1; ES2349911B1

Abstract

Methods and systems for analysing sealing in fluid deposits, that include the following steps: a) injecting a tracer gas into the deposit at a predetermined pressure and temperature; b) capturing external images of the deposit using an IR camera in a predetermined zone of the spectral band in which said tracer gas is optically active; c) locating unintended holes in the deposit using gas leaks detected by displaying said images or images obtained from them. The systems include an IR camera fitted with a filter for capturing images of the outside of the deposit and a computer connected to it provided with a software program for displaying the images captured by the camera or processed images based on them. The invention also relates to a computer program for performing the method.

Description

METHODS AND SYSTEMS FOR PERFORMING SEALING ANALYSIS IN FLUID TANKS

FIELD OF THE INVENTION

The present invention relates to methods and systems for performing leakage analysis in fluid tanks and more particularly in aircraft fuel tanks.

BACKGROUND

When you want to get deposits with a guarantee of tightness, it is necessary to control and identify possible leaks that may occur due to manufacturing defects or pores in the materials used, especially if they are fuel tanks in airplanes, so it is essential to count with appropriate methods for the detection of possible pores or fissures in the deposits.

In the known technique, the procedure used to analyze the tightness of aircraft fuel tanks consists in applying, first, pressure to the air contained therein and subsequently analyzing possible pressure losses that may occur due to the existence of pores and fissures and, secondly, apply a coating of an appropriate composition (for example, soapy water) to the tank to detect leakage points.

In that specific aspect, as in many others, the industry demands improved methods and the present invention is oriented to the satisfaction of that demand by applying techniques based on infrared spectral image (IR) as considered in the following following publications. - S. Briz, from Castro AJ. , López F., and Scháfer K., "Remote Sensing of

Ozone by Open-Path FTIR Spectroscopy. Analysis and Validation of Different Analysis Techniques "Proc. Of Chemical Industry and Environment IV. VoI. 2. A. Macías & J. Umbría Eds., 2003, pp 313-323, Spain.

- J. M. Aranda, S. Briz, J. Meléndez, A. J. de Castro, F. López, "Fíame analysis by IR thermography and IR hyperspectral imaging". Proc. of Quantitative Infrared Thermography V-QIRT2000, 337-342. (2000)

- S. Briz, AJ. de Castro, F. López "Modify resolution study of the ground-based passive emission of ozone" Applied Optics, 39, 1980-1988 (2000)

- AJ. de Castro, J. Meneses, S. Briz, F. López "Non dispersive infrared monitoring of NO emissions in exhaust gases of vehicles" Rev. Sci. Instrum., 70, 3156-3159 (1999)

- J. Meneses, S. Briz, AJ. de Castro, J. Meléndez and F. López. "New spectral selection system for infrared imaging of carbon monoxide in combustion environments". Combustion Diagnostics. M. Tacke Ed. SPIE Proc. 3106, 105 (1997)

SUMMARY OF THE INVENTION

An object of the present invention is to provide methods and systems for the detection of leakage faults in fluid storage tanks that can be implemented both in their place of manufacture and in their use.

Another object of the present invention is to provide non-intrusive methods and systems for detecting leakage failures in fluid storage tanks and, particularly, in aircraft fuel tanks.

Another object of the present invention is to provide methods and systems for real-time detection of leakage failures in fluid storage tanks and, particularly, in aircraft fuel tanks. In a first aspect, these and other objects are achieved with a method of analyzing the tightness of a reservoir for the storage of fluids, which comprises the following steps: a) introduce a trace gas at a predetermined pressure and temperature into the tank; b) take external images of the tank by means of an IR camera in a predetermined strip of the spectral band in which said trace gas is optically active; c) locate pores or manufacturing defects of the tank from the gas leaks detected by means of the visualization of said images or of images obtained from them.

In a second aspect, these and other objects are achieved with a system for detecting leakage faults in fluid storage tanks by detecting leaks of a trace gas introduced in said tank, comprising: a) an IR camera equipped of a filter for the taking of images outside the tank in a predetermined strip of the spectral band in which said trace gas is optically active; b) a computer connected to said IR camera provided with a "software" that allows to locate pores or manufacturing defects of the deposit by displaying the images taken by the camera or images obtained from them by one or more of the following processes : b1) a process of optimizing its contrast; b2) a comparison process that allows to identify variations between images taken at different time points.

In the methods and systems according to the present invention, both the IR absorption capacity and the IR emission capacity of the trace gas can be used by setting, in each case, the corresponding temperatures for the trace gas introduced into the tank and to adapt the tank as IR source

In a preferred embodiment of the present invention, the leak detection is carried out in stages directed specifically to the location of large leaks, small leaks and very small or intermittent leaks, using, respectively, the images taken by the IR camera, images obtained applying to the images taken by the IR camera a process of contrast optimization and images obtained by processing Temporary of the previous images. This optimizes the leak detection process using specific means for each type of leak.

In another preferred embodiment of the present invention, images obtained by applying to the images taken by the IR camera a contrast optimization process consisting of a gain control process selected from those of linear type are used for the detection of small leaks , logarithmic, exponential, power of 2, square root, power of 3 and 1/3 and with the typical methods of dynamic range organization (RD): "Full RD" and "select percentage of the RD". It is thus facilitated that the user of the system has appropriate images to detect leaks in each area of the tank according to the level of resolution required.

In another preferred embodiment of the present invention, images obtained by comparing processes of images of different temporal moments are used for the detection of intermittent leaks and very small leaks in which the temporal separation between the compared images and the scaling of images can be selected background of the image resulting from the comparison. It is thus facilitated that the user of the system has appropriate images for the detection of such leaks.

In another preferred embodiment, the fluid reservoir is a fuel tank of an airplane and the trace gas is CO2. This achieves a method and a system of analysis of the tightness applicable in an environment such as that of an aircraft assembly plant since CO2 is an innocuous gas since its possible emission to the environment due to eventual leaks from the tank does not affect at its usual concentration more than what other usual activities do in that plant. On the other hand, there is no visual pollution since CO2 is a transparent gas in the visible spectrum, so it does not affect in any way other methods that are developed in the factory that use said spectral band and, in turn, CO2 is an inert gas that does not stain or contaminate any surface or gap by direct contact. One of the fundamental characteristics of the system object of the present invention is that of being an image system, even of vision, since it allows to detect phenomena by means of the image process and to take decisions based on quantitative parameters on them. The other fundamental characteristic is that said image has a spectral property: it does not work as a standard image system in the band provided by the camera manufacturer but only collects signal in the spectral band that optimizes the contrast that can be changed based on variables relevant to the effect such as the trace gas used, the temperature, the IR scenario on which one works. That is its spectral character: a characteristic wavelength of contrast optimization to optimize the detection of the leak, especially the very small ones. Other features and advantages of the present invention will be apparent from the detailed description that follows in relation to the accompanying figures.

DESCRIPTION OF THE FIGURES

Figure 1a shows a photo of a fuel tank sector of an airplane and Figure 1 b the corresponding IR image after optical filtering.

Figures 2a and 2b show the IR image processed with, respectively, logarithmic stretching and exponential stretching. Figures 3a and 3b show respectively the IR image processed with, respectively, linear "stretching" in "determined dynamic range" mode, and linear "stretching" in "90% RD" mode.

Figures 4a and 4b show two images of a sequence in which the effect of a leak can be seen. Figure 5 shows a block diagram of the "software" used in the present invention.

Figure 6 shows the main window of the interface of said "software" displayed.

Figure 7 shows the "Histogram of the application" tab of the interface of said "software". DETAILED DESCRIPTION OF THE INVENTION Method

We will first describe a preferred embodiment of a method according to the invention for the analysis of the tightness of a fuel tank of an airplane, using the IR absorption capacity of the trace gas. The method comprises the following steps: a) The tank is filled with CO2 as trace gas at a predetermined pressure and temperatures depending on the type of tank and the environmental conditions. These variables can be used to optimize the detection, depending on the conditions and the type of leakage. b) In order to increase the contrast and optimize the detection, the tank itself is used as an active IR source, for this purpose it is heated using a thermal source, such as an air heater or an IR lamp, until a optimum temperature for the contrast that, although it can vary depending on the temperature of the environment or the light environment, is in normal conditions between 3O ⁰ C and 5O ⁰ C. In the area selected for the analysis of the tightness the Ia is controlled temperature by using an IR pyrometer. c) Next, the desired area of the deposit is inspected using an IR camera to which the spectral detection range has been modified, to optimize the gas-bottom contrast in the spectral band in which the CO2 is optically active .

Simultaneously to the taking of images, and for very small leaks, it is advisable to "sweep" the ambient CO2, for example using a standard gun connected to a bottle of an IR transparent gas, such as dry N2, whose jet is directed in The direction of the optical path of the measurement, thus evacuating an important part of the ambient CO2, which would prevent the masking of the presence of trace gas from the leak, and therefore its detection in the inspection areas. In addition, the movement of trace gas caused by nitrogen favors the detection of trace gas, both in direct vision and with the proposed temporary processing. The elimination of atmospheric CO2 interference can also be achieved by covering and sealing the area to be inspected with a plastic transparent to the radiation characteristic of the trace gas, in order to create a nitrogen atmosphere with which the leaks of CO2

In the specific embodiment that we are describing, the imaging is done by an IR camera provided with an IR optical filter, which is coupled to the lens through an internal filter holder designed for this purpose and anodized to avoid internal reflections, centered and with the Optimized bandwidth for the detection of trace gas, always in the environment of the center of the CO2 absorption band. Both centering and width are carefully determined for an optimized detection by means of the figure of merit developed specifically called: Relative difference of detected spectral Radiance. In essence, this figure evaluates for a given work and deposit conditions, the difference in optical signals (radiance) that would be detected, by an IR detector, in the presence and absence of leakage gas in a given environment, normalized to the signal IR background in the field of vision of the system. It is essentially defined as (| ΔL | / L) and has no units. Optimizing the system means increasing the value of this function for parameters that can be freely modified such as the bandwidth of the filter and its centering, which are the ones that will optimize the detection, in particular in the case of small leaks.

The selection of the trace gas and the determination of the strip of the spectral band in which the IR images must be taken is an important aspect of the present invention. It is necessary to take into account in this regard that it is a question of carrying out a tightness analysis of fluid deposits and, in particular, of fuel tanks of airplanes in their manufacturing plants, that is, in places where other activities are carried out. and where there are certain environmental conditions.

In principle, any polar molecule gas would be appropriate for the present invention in that it has an infrared signature that can be uniquely detected. However, taking into account the environmental conditions of an aircraft assembly plant, the CO2 Another gas that could be appropriate for such environmental conditions, although it has some disadvantages such as a huge greenhouse effect (which would limit its application to special cases) is sulfur hexafluoride, for which a different spectral band should be selected in the Thermal IR.

The IR camera is mounted on a mobile tripod to ensure stability, precise positioning and flexibility for its movement during imaging and is located at a distance of around 50 cm from the sector to be inspected, which is framed by selecting the height and adequate angle, although this distance may change if the optics or the dimensions of the area to be inspected are changed. d) Detect the position of the pore or defect that produces the trace CO2 leaks by displaying the IR images taken by said IR camera and / or images derived from them by applying certain processes. Leak detection is performed according to a leak discovery procedure that goes from the largest to the smallest. This means that it starts with fast and comfortable modes for the detection of large leaks, sweeping most of the tank to be analyzed, to slower and more expensive ways to detect the smaller ones, which will be limited to more restricted areas thereof, where the Experience and previous analysis make the small leak (valves, joints, "bondings", ...) more likely.

Therefore, the analysis of each deposit will be carried out in the specified order, the modes mentioned below, so that it would be passed from one mode to the next only if no leaks are detected through it and, if it has also been detected The existence of leaks by global, non-discriminatory procedures. For this, the classic procedure of the measurement of the rate of fall of the total pressure in the pressurized tank, indicated above (measuring the rate of fall of the overall pressure) is used.

A first way of detecting and locating larger leaks, following the method object of the present invention, is the direct visualization of the IR image that supplies the IR camera thanks to the spectral optical filtering performed. The advantage lies in the rapidity of the analysis. This filtering consists in selecting a spectral range of the working range of the IR chamber in which the selected trace gas (CO2) is optically active in order to increase the signal-to-bottom ratio and optimize the detection of the trace gas. Thanks to this, the detection of many of the leaks by direct vision of the IR image is allowed. In addition, the filtering characteristics can be optimized, by "tuning in" the filter, both centered and bandwidth, depending on the measurement conditions, distance, ambient temperature, gas temperature and the background of the scene, size of the leak or Other relevant parameters. In Figure 1 a you can see a photo of an area of a fuel tank located in the wing of an airplane with a filling pipe and in Figure 1 b the corresponding IR image obtained after the selected IR optical filtering to optimize the detection of trace gas (CO2). In the first approximation and assuming that the temperature of the gas expelled by the leak is of the order of the environment, the variation of the digital level of a pixel is proportional to the variation of the radiance of the corresponding point of the bottom of the scene multiplied by the transmittance of the optical path Therefore, the extra presence of gas in the optical path due to the leak will modify the radiance, decreasing (if the gas escapes at a lower temperature or of the order of the environment) the digital level of the affected pixel, with respect to the rest. This provides a contrast in the image between the pixels affected by the leak and those that are not, which, supported by the visual acuity of the operator, is in many cases sufficient to detect the leak directly.

A second mode of detection and location of leaks following the method object of the present invention, indicated for smaller leaks in accordance with the aforementioned, is a processing of the IR images supplied by the IR camera in which an automatic control of the gain that allows to optimize the visualization of the sequence of images, facilitating the detection of possible leaks. As noted, the detection of the presence of trace gas is due to the absorption of part of the energy from the IR source, so that for small leaks the difference in levels will not be very high. It is therefore that images provided by the IR imaging device must be properly displayed.

Among the automatic gain control methods available, you must choose the appropriate one for each case. In this sense, the results obtained with different stretching methods are shown in the same image of Figure 1 a.

Figure 2a shows the IR image with logarithmic stretching. It can be seen that the dark areas of the original image appear more clearly represented, while in the light areas detail is lost.

Figure 2b shows the IR image with exponential "stretching", and it can be seen that the opposite occurs since the light areas of the scene are enhanced in front of the dark ones.

Figure 3a shows the IR image to which the linear "stretching" has been applied in a "determined RD" mode, specifying the minimum and maximum values of the display range. It can be seen that the dark areas of the image are visualized in greater detail, and that the contrast in certain areas has increased with respect to the original image. You can see a slight movement in the sequence due to the leakage. Figure 3b shows the IR image to which it has been applied.

linear "stretching" in "90% Dynamic Range" mode. Again, the visualization of the leak is very weak, although more appreciable than in the image

IR without automatic gain control.

These different automatic gain methods can be applied to a selected area of the image if this is of interest.

A third mode of detection and location of leaks, indicated for leaks that could not be detected by the previous procedures, following the method object of the present invention, consists of a temporary processing based on the detection of temporary variations in the image of reference caused by the trace gas when it diffuses into the atmosphere when leaving through a fissure or pore of the fuel tank. It is very effective for detecting leaks in very unfavorable situations: when the pores are Small or leaks are intermittent. In this method, a first dynamically adjustable parameter that allows highlighting the temporal variation of the processed image against the reference image and a second dynamically adjustable parameter that allows choosing the temporal separation between the reference image and the one to be processed are highlighted

Figures 4a and 4b show two images of a sequence in which the effect of a leak can be seen. The area of the image where the trace gas appears is darker since said gas absorbs the IR radiation. In both images the effect of a trace gas leak through a fissure can be observed. The different appearance of both images is due to the real-time selection of different processing and stretching parameters, in order to optimize the visualization of the leak. The second image, due to the greater contrast, allows the leak to be detected much more clearly, as well as to locate its origin. On the other hand, the visualization of a complete sequence of images facilitates the detection of leaks since the movement of the trace gas is clearly distinguished on the static background of the sequence.

An important aspect of all modes of operation is to determine a good level of system zero, prior to the leak detection process. The level of "zero" would be the digital level that would provide a pixel that, belonging to the scene of the tank to be analyzed, is not affected by leaks and that is in the average environmental conditions and distance, for this it is enough to focus on an area Let it be known that there is no leak. This level will be used to obtain the parameters on which the camera should initially be centered in order to obtain the greatest possible dynamic range. The level of zero must be repeated from time to time and especially if changes are observed in environmental conditions, observation, background, etc.

System

We will now describe a preferred embodiment of the system object of the present invention for the analysis of the tightness of a fuel tank of an airplane, using the IR absorption capacity of the trace gas introduced into the tank. The system includes: - An IR camera in the middle IR band (3-5 microns) as described above optimized spectrally for the detection of CO2 by means of a filter.

- A laptop with a central processing unit and peripheral devices appropriately sized.

- A "software" that, as shown in Figure 5, consists of three main modules:

- An image acquisition module 11 with a sub-module 13 for acquiring IR images from the IR camera and a sub-module 15 for acquiring images from a file.

- An image processing module 21.

- An image storage module 31.

In the preferred embodiment that we are describing said "software" has been developed under the "LABVIEW" platform of National Instruments, but as the expert in the field will well understand, it could have been developed over any other.

The "software" allows a first mode of operation in real time in which the IR images are acquired, processed and displayed in real time and a second mode in which the aforementioned functions are applied to previously stored sequences of IR images.

The "software" has a visual interface that allows modifying in real time the parameters of acquisition, visualization, processing and storage.

There are also numerous indicators that show relevant information.

The "software" interface consists of several windows and tabs, in which the controls and indicators related to each functionality are grouped.

Figure 6 shows the main window 41 of the application that allows the control of the acquisition, visualization, processing and storage of images in file. The other General Information window 43 collects additional "software" configuration parameters. The main window contains two display screens showing the original IR images and the processed images and three tabs: - Processing 51 (displayed in Figure 6): It allows selecting the acquisition and processing parameters.

- Histogram 53: It allows to visualize the histogram of the original and processed IR image, and to select the correction parameters of the erroneous pixels. - Recording 55: Contains the storage parameters of the original and / or processed sequence in standard formats, both uncompressed (FITS or similar) and compressed (AVI, MPEG, or similar).

The main window also contains several control buttons: START / STOP, STOP, Help. It also contains Saturation and ONLINE indicator buttons for easy handling. For example, the Saturation indicator that appears in the IR camera acquisition mode is activated if the gray level of any pixel in the image exceeds the maximum value selected in the Maximum value received control, in the General Information window. Image acquisition module 11 Image acquisition module 11 generates a sequence of IR images in a format suitable for processing. This sequence can come from a previously stored file or be acquired directly from an IR camera connected in real time. This module differs slightly in both acquisition modes and its main functions are explained below. In both cases, the desired processing and visualization speed must be established, and an image flow is provided at the output at the specified rate and in a format suitable for the following modules.

Since the main objective of the system is the detection of leaks by means of the visualization of the trace gas, one of the capacities of the system object of this invention is that of varying the speed of visualization to optimize that task. If the escape of gas through a fissure occurs slowly, the variations observed in the image will be small, and it will be preferable to reduce the number of images processed per second to be able to appreciate the variation in the sequence and help the visual detection of the leakage. On the contrary, if there are rapid variations in the sequence, it will be necessary to increase the acquisition rate to facilitate the detection. Of this Thus, the speed of the system can be adjusted according to the output speed and the movement of the trace gas through the leak, optimizing the visualization of changes by the observer.

Therefore, an important parameter of the "software" that must be properly selected is the acquisition speed of the camera in images or "frames" per second (fps). The "software" allows you to select an arbitrary fps rate per second, so that it can be used with any camera, without limiting its acquisition speed. The final speed of the system can be specifically defined depending on the type of leakage expected, so that the camera must have adequate performance in terms of acquisition speed. In that case, the number of images processed per second delivered by the "software" will be a function of several parameters:

- The speed of supply of images of the camera, which is characteristic of the camera itself and cannot be exceeded. Therefore, this is a parameter to consider when acquiring the camera.

- The processor used for the execution of the program and its computational load.

- The desired rate of images per second, which can be selected by the user depending on the type of leakage and gas output rate, as indicated above, as well as the observation conditions and IR brightness of the scene. Although the user can select an arbitrarily high acquisition rate, if it exceeds the limit imposed by any of the above parameters, the effective exit rate will decrease. The user can detect this situation by observing the FPS output indicator, which indicates the actual processing rate. If it does not reach the speed selected in the FPS control, it may be convenient to reduce the latter.

The "software" can also acquire images from uncompressed files which, in the preferred embodiment we are describing, are of FITS and ABMOV formats. The FITS or FTS format ("Flexible Image Transpon System ') is the format selected for storing IR images for further processing. This is a standard format for storage of images without compression, with a header that includes information of the IR image, and also allows adding optional fields with specific data of the measurement, the day, the processing parameters used, etc. The ABMOV format is the non-compressed image storage format used by the IR camera that has been used, and that may be different for another camera.

In this operating mode, unlike the camera acquisition mode, the "image jump" control appears, which allows you to process as many images as the control indicates, in order to speed up the processing and more easily detect slow variations in The sequence. The default value of this control is 2.

Image Processing Module 21

This module receives at the input the sequence of images generated by the acquisition module 11, and provides at its output the sequence of images processed in FITS format for later storage by module 31.

A series of operations are carried out in the processing module, in order to optimize the sequence display and facilitate leak detection. The most relevant are summarized below. Automatic gain control and elimination of anomalous pixels

The "software" allows simultaneous viewing of the original and processed sequence. In order to be able to see in detail certain regions and, in this way, to be able to discern more easily the reflections and possible hidden elements (for example, located in very dark or very bright areas) of possible leaks, a function known as automatic control of gain, "stretching" or adaptation of the dynamic range.

In addition, the "software" allows the adaptation of the RD to a "region of interest" (RDI) within the image, selected at the request of the user when he suspects the existence of a small leak in a certain area, much lower than the total area displayed .

The "stretching" aims to highlight or differentiate more clearly certain regions of the image. The "software" allows to apply the control automatic gain to the entire image or to an area of interest selected by the user. The controls for selecting the "stretching" parameters appear in the Processed tab 51 of the main window 41.

The types of "stretching" that allows us to apply the "software" that we are describing are: linear, logarithmic, exponential, power of 2, square root, power of 3 and 1/3. The use of one type or another is determined by the region of the dynamic range that you want to visualize in greater detail. For example, the logarithm function enhances the low values of the digital level against the high ones, allowing a thorough analysis of the dark areas of the image. On the contrary, the exponential function makes it possible to discriminate more clearly in the clear regions of the image. The linear stretching also weighs high and low digital level values without altering the input image.

The "stretching" methods that allow us to apply the "software" that we are describing are "RD complete", "Selectable percentage of RD". The "stretching" method determines the region of the dynamic range to which automatic gain control is to be applied. Since images are obtained from the camera with a resolution greater than the one that can be displayed, the "stretching" method allows selecting the way to adapt to the display screen the dynamic range of the image obtained from the camera, highlighting certain areas of The image

The "full RD" method allows the visualization of the image throughout its dynamic range, between its maximum and minimum values. It is a dynamic method that does not allow the user to specify any parameter related to the visualization. Since the extreme values of the image are used to determine the full dynamic range of the image, the presence of anomalous pixels with extreme values can negatively affect the visualization. In this mode, the Max and Min controls are disabled and indicate the extreme values of the pixels of the image, which may be useful for detecting the presence of anomalous pixels. The "determined RD" method is similar to the "full RD", except that it allows the user to define the minimum and maximum values of the dynamic range of display. This method can be used to increase the contrast of certain regions of the image.

The "90% RD" method dynamically eliminates the lower and upper 5% of the dynamic range of the image, thus displaying the range from 5 to 95% of the values. This method is more robust than "full RD" since it is not sensitive to specific errors in the image. It is recommended to use this method if the presence of anomalous pixels is detected by observing the histogram of the image or the maximum and minimum value indicators. Like "RD complete", it is a dynamic method that does not allow the selection of the extreme values of the visualization, but shows them in the Min and Max indicators.

The method "given percentage of the RD" is similar to the previous one except for the possibility of selecting the percentage dynamic range of visualization. In this case, in the Min% and Max% controls, the percentage of pixels with the lowest and highest digital value, respectively, is specified, which will not be displayed.

The "software" allows the application of automatic gain control to both the original and the processed image independently, being possible to select different types and methods of "stretching" to visualize both sequences, since they will generally have different characteristics. The method used by default is "RD complete", both for the original image and for the processed one.

It is convenient to combine the automatic gain control with the elimination of anomalous pixels to optimize the display of a certain sequence. For this, it may be interesting to observe the histogram of the original and processed image, in which the frequency of the digital levels is represented along the dynamic range, that is, the number of pixels of each digital level of the image. The histogram of an image therefore provides an intuitive idea of its range of values, and allows the appropriate selection of the most appropriate type and method of stretching. The correction of anomalous pixels aims to eliminate the effect of those pixels that, after processing and due to the mathematical operations applied to the image, take anomalous extreme values that distort the dynamic range of the image, seriously affecting the visualization.

To detect the existence of these anomalous pixels, the histogram of the processed image must be observed and see if the digital levels are distributed throughout the dynamic range presenting the shape of a Gaussian distribution or if on the contrary there are extreme values. On the other hand, it is also possible to detect said anomalous pixels by observing the maximum and minimum value of the image in the Max and Min indicators of the automatic gain control panel in "full RD" or "determined RD" mode. Figure 7 shows the tab 53 Histogram of the application, which allows viewing the histogram of both the original IR image and the processed one. In addition, in this tab you must select the parameters for deletion of the wrong pixels.

As already mentioned, an alternative way of avoiding the effect produced by these pixels of extreme values is to apply automatic gain control by specifying the maximum and minimum values of the display, or the percentage of pixels eliminated at the ends. In addition, it is possible to select a region of interest in the area that is desired to be visualized avoiding said anomalous pixel. Temporary processing

In the original IR image, the value of each pixel is proportional to the radiance detected. The proposed algorithm is based on performing a transformation to obtain an image where the value of each pixel is proportional to the variation of the detected radiance of that pixel in the anterior and posterior instants.

In this sense, an important aspect of the present invention lies in the use of the properties of vision systems based on their synergy with the operator. The system object of the invention converts the detection of fuel leaks from a tank in the detection of slight variations of the scene. In this way, the detection of something static: pore or defect, becomes something dynamic: gases moving in its environment, facilitating its detection by means of the IR camera - operator system. With this enhances the user's ability, based on their own training and experience, to generate skills that allow, for each type of leakage, of the IR fund and other operating conditions, to make the decision about what is the most appropriate value of the mentioned parameters for the detection of a small leak in a given situation. Storage Module 31

This last module of the application allows the storage in memory of the original or processed sequences for later viewing or reprocessing. The application offers different storage options depending on the mode of operation, which are described below.

In the IR camera acquisition mode, it is possible to store both the original and the processed image in FITS format. There are a number of parameters that must be properly selected before starting the recording. The application allows you to start recording at any time, and can be configured in various ways depending on the type of leakage. It is possible to generate a single continuous sequence of the specified length, or on the contrary a series of sequences separated by a waiting interval, which will allow detecting intermittent leaks without the need to store an excessively large number of images.

In the acquisition mode from file, the storage of the processed sequence in the FITS, AVI or both formats is allowed simultaneously.

Among the advantageous characteristics of the method and system object of the present invention, the following may be indicated.

• Remote detection and location. It is a non-intrusive measure that allows optical methods to detect and locate where a leak occurs.

• Versatility in spatial resolution. Simultaneous detection with a single sensor system in a wide area of space. Depending on the chosen optics, it can be achieved from sweeping a wide area with low resolution to a small area with high resolution. • Instant temporary response. The detection takes place in real time and, in addition, it is possible to carry out a subsequent study of verification and control from the processing of previously stored images.

• Complete transportable system. The proposed IR imaging system consists of a high-performance IR camera, as well as a "software" that can be installed on a portable computer that allows the control, acquisition and real-time processing of images from the IR camera. This system allows its transfer and use outside the factory.

• It is a flexible and adaptive "vision system" that has the ability to incorporate training and user experience into decision-making, to generate skills that allow optimization of the detection for each type of leak and depending on the IR background and other operating conditions.

Although the present invention has been described entirely in connection with preferred embodiments, it is evident that those modifications can be made within the scope of, not considering this as limited by the previous embodiments, the following claims.

Claims

1.- Method of analyzing the tightness of a reservoir for the storage of fluids, characterized in that it comprises the following steps: a) introducing a trace gas into the tank at a predetermined pressure and temperature; b) take external images of the tank by means of an IR camera in a predetermined strip of the spectral band in which said trace gas is optically active; c) locate pores or manufacturing defects of the tank from the gas leaks detected by means of the visualization of said images or of images obtained from them.

2. Method of analysis of the tightness of a tank according to claim 1, characterized in that said step c) comprises one or more of the following steps: d) detection of larger gas leaks by direct visualization of the images taken by the IR camera; c2) detection of gas leaks of smaller size by displaying images obtained by subjecting the images taken by the IR camera to a process in which its contrast is optimized; c3) detection of very small or intermittent gas leaks by displaying the images obtained by subjecting the images taken by the IR camera to a comparison process that allows identifying temporary variations in them.

3.- Method of analysis of the tightness of a tank according to claim 2, characterized in that the process of optimization of the contrast of stage c2) is a gain control process selected from those of linear, logarithmic, exponential type, power of 2, square root, power of 3 and 1/3 and those of the method "Full Dynamic Range", "Determined RD", "90% of the RD" and "RD of a certain percentage".

4. Method of analysis of the tightness of a tank according to any of claims 2-3, characterized in that the process of comparing images of stage c3) includes the selection of the temporal separation between the compared images and the background scaling of the image resulting from the comparison.

5. Method of analyzing the tightness of a tank according to any of claims 1 -4, characterized in that said step a) comprises a stage of adaptation of the tank as an IR source.

6. Method of analysis of the tightness of a tank according to any of claims 1 -5, characterized in that:

- the IR absorption capacity of the trace gas is used;

- the trace gas is introduced into the tank at a temperature lower than the ambient temperature;

- the tank is adapted as an IR source by heating it to a temperature higher than the ambient temperature.

7. Method of analysis of the tightness of a tank according to any of claims 1 -5, characterized in that:

- the IR emission capacity of the trace gas is used; - the trace gas is introduced into the tank at a temperature higher than the ambient temperature;

- the tank is adapted as an IR source by cooling it to a temperature below room temperature.

8. Method of analyzing the tightness of a tank according to any of claims 1-7, characterized in that said tank is a fuel tank of an airplane.

9. Method of analysis of the tightness of a tank according to any of claims 1-8, characterized in that said trace gas is CO2.

10.- System for analyzing the tightness of a reservoir for the storage of fluids by means of the detection of leaks of a trace gas introduced in said reservoir, characterized in that it comprises: a) an IR camera equipped with a filter for the taking of images from outside of the deposit in a predetermined strip of the spectral band in which said trace gas is optically active; b) a computer connected to said IR camera provided with a "software" to locate pores or manufacturing defects of the deposit by displaying the images taken by the camera or images obtained by one or more of the following processes: b1) a process optimization of its contrast; b2) a comparison process that allows to identify variations between images taken at different time points.

11. System for analyzing the tightness of a tank according to claim 10, characterized in that said tank is a fuel tank of an aircraft.

12. System for analyzing the tightness of a tank according to any of claims 10-11, characterized in that said trace gas is CO2.

13. A computer program adapted to execute the method of any of claims 1 to 9.