DE102010019399B4 - Image acquisition method for IR images and thermal imaging camera - Google Patents

Abstract

Image recording method for IR images (9, 10, 11, 12, 13), wherein for different focus settings (14, 15, 16, 17, 18) IR images (9, 10, 11, 12, 13) are recorded, wherein from the recorded IR images (9, 10, 11, 12, 13) for a given image area with a base transformation transformed images are calculated, wherein derived from the transformed images in each case a measure (20) and the associated focus setting (14, 15, 16, 17, 18), the measure (20) being monotonically dependent on the high frequency portion in the transformed image and on a number of edges in the IR image (9, 10, 11, 12, 13) from the derived and associated measures (20) an optimal measure (24) is determined, wherein one of the optimal measure (24) corresponding focus setting (27) is set and wherein when determining the optimal measure (24) an interpolation curve (23) the derived measures (20) as a function of zugeo rdneten focus settings (21) is calculated.

Description

The invention relates to an image recording method for IR images and a thermal imaging camera with a recording device for IR images, wherein the receiving device has an adjustable focusing device.

In order to increase the range of use of thermal imaging cameras and to improve the quality of the recorded IR images, it is known to equip the thermal imaging cameras with a manually operated or motor-assisted, manual focusing unit in order to be able to image objects arranged at different distances as sharply as possible. Since the known solutions so far allow only very inaccurate focussing, they have found little practical use.

From the US 5,404,013 A For example, there is known an infrared imaging system with automatic focus control in which, for a selected image line, an associated analog signal is electronically processed to adjust a speed control logic of a motor for focus.

From the DE 696 286 54 T2 For example, a measuring system for determining the global modulation transfer function is known, in which the modulation transfer function of an optical microscope is determined by means of a two-dimensional test pattern.

From the US 5,075,713 A a focus system of a camera apparatus is known, which has the possibility of automatic and manual control, wherein a front lens group and a master lens group are movable, wherein a natural operating feeling is achieved by the master lens group in both the automatic control mode and in the adjustment ranges in the manual control mode is slidable to enable a focus over a wide range.

From the US 2006/0109369 A1 An autofocus system is known which contains a rangefinder in addition to a contrast autofocus. With the rangefinder, the distance to the object is monitored and only when a major change occurs a readjustment using the contrast autofocus performed.

The invention has for its object to improve the performance characteristics of a thermal imaging camera.

To solve this problem, the invention provides the features of claim 1 in an image recording method for IR images. In particular, it is thus provided that IR images are recorded for different focus settings, that a measurement number is derived from the recorded IR images for a given image area and assigned to the associated focus setting, wherein the measurement number is monotone of the proportion of high frequencies in the transformed image and depending on the number of edges in the IR image, an optimal measure is determined from the derived and associated measures, and a focus adjustment corresponding to the optimum measure is set. In this context, basic transformations are generally understood to mean transformation rules in which the coefficients of a decomposition of the respectively examined IR image with respect to a function basis are determined. A monotonic dependence is understood to mean a functional dependency in which, in the case of increasing monotonicity, the functional values at least do not fall or even increase with increasing argument and in which case the function values do not rise or even fall with increasing argument in the case of falling monotony. Thus, rising and falling monotony are reversed in their behavior to each other.

The invention thus has the advantage that in an automated process by the evaluation of the recorded IR images, an optimal focus adjustment is determined and adjusted, for which an optimal focus is achieved. The user of the image capture process is therefore exempt from self-determination of this focus by manual or motorized adjustments. The performance characteristics of a thermal imaging camera with the image sensing method are therefore considerably improved.

Preferably, the measure serves as a measure of the sharpness. This sharpness dimension can thus be defined as allowing an objective rating of each IR image with respect to the number of high frequencies or edges in the image. Here, a higher number of edges indicates a greater sharpness than a smaller one.

For the automated implementation of the method, it may be provided that the focus adjustment is changed with a monotonous sequence of distance settings until the sequence of assigned measures has an extreme value.

It can be provided that in each case a relative or absolute adjustment path is read out on the objective relative to the focus settings of the sequence. The advantage here is that the read-out adjustment paths are available for starting the optimal focus setting. It is thus not required that a fixed relationship between focus settings and adjustment paths is deposited.

It is advantageous in the invention that the described method does not have to start from a defined focus adjustment and that the knowledge of the correspondence between focus adjustment and an absolute adjustment angle on a lens need not be known. Rather, the method can start from any focus setting, and it is sufficient to set the determined optimal focus setting already the knowledge or specification of relative adjustment angles or adjustment paths on the lens.

In order to easily determine the extreme value of the measure, it can be provided that the focus adjustment in a first step, first with a monotonous sequence of distance settings with a first monotony, for example, rising or falling, is changed and in a second step with a monotonous sequence of distance settings is changed with respect to the first monotony reversed monotony (falling or rising) as soon as the sequence of the assigned measures deviates in the first step from a predetermined monotony.

In order to arrive at the optimal focus setting particularly quickly, that is to say with particularly few IR images, provision can be made for the focus adjustment to be changed by a step width which is determined by the direction and / or the magnitude of the change in the associated numerical values. In particular, it can be provided here that the step size in the vicinity of the optimum measure is selected to be small, while it is set large at a great distance from the optimum measure.

An embodiment that is already sufficient for many applications and that can be implemented in a simple manner can provide that when deriving the measured value from a transformed image, an average value of the magnitudes of the transformed image is formed in a frequency range. Preferably, a frequency range in the upper half or even in the upper quarter of the average or typical frequency range of an IR image is chosen.

It can also be provided that, for the calculation of the transformed image, a feature analysis or extraction is carried out as base transformation. For example, edges can be detected in the IR images.

In order to be able to find the maximum or minimum of the sharpness function without completely slowing through the range of the focus settings, the ratio between the signal and the noise of the metric, wherein the noise is caused by image noise, must not be too low. It has been found that excessive noise ultimately leads to local extremes in the metric that affect the global extreme value search. If scenes with relatively lower contrast of, for example, about 5 ° C. are to be displayed with the method according to the invention, it may be advantageous if the recorded IR images and / or the transformed images are smoothed before deriving the measure, for example with a simple mean value filter ,

In order to obtain the best possible focus adjustment even with a relatively large step size of the adjusted focus settings, it can be provided that an interpolation curve to the derived measures is calculated as a function of the assigned focus settings when determining the optimal measure.

It can be provided that a predetermined focus setting is selected if the sequence of the derived and associated measures does not allow the determination of an optimal measure. The advantage here is that a typical focus adjustment, which may correspond, for example, a fixed focus, is selected when the recording situation is unfavorable for the functioning of the rest of the image recording process.

The mathematical structure of the base transformation used is selected according to the invention according to the present recording conditions of the thermal imaging camera or the current recording situation. For example, it can be provided that the transformed images are calculated using a 2D Fourier transformation, in particular with an FFT (fast Fourier transform) or with a DFT (discrete Fourier transform). It may alternatively or additionally be provided that the transformed images are calculated with a base transformation with a base of wavelets and / or with a cosine transformation.

It is particularly favorable if the focus settings are respectively designated by a distance setting of an objective, the images being recorded by the objective. The advantage here is that even conclusions about the distance to the imaged object are made possible by the determined focus setting.

For a manual tracking or fine adjustment, for example in complex scenes, it can be provided that a control connection for controlling the focus settings is disconnected or disengaged, in particular mechanically and / or electrically and / or electronically, after which a focus adjustment is manually adjusted.

To solve the problem, the invention provides the features of claim 9 in a thermal imaging camera of the type mentioned. In particular, it is provided that an image processing unit is formed, which is set up to calculate a base transformation of a predetermined image area of a captured IR image into a transformed image, that an evaluation unit is designed, which is adapted to calculate a measure for the respective transformed image, wherein the metric is monotonically dependent on the proportion of high frequencies in the respective transformed image and the number of edges in the IR image, and that a control unit is provided, with which the focussing unit can be controlled so that the calculated to the resulting IR image Measure number assumes an optimal value. Preferably, the device of the image processing unit, the evaluation unit and / or the control unit by appropriate programming of the operating program of a data processing unit, which is integrated into the thermal imager.

It can be provided that the data processing unit is programmed in such a way that it is set up to execute a method according to the invention.

In order to enable a correction of the focus, for example, in complex scenes or unfavorable shooting conditions, it can be provided that a driving connection between the focussing unit and the control unit off and is engageable, with a disengaged control connection, the focussing unit is manually operated or is actuated. The advantage here is that the inventive method requires no defined starting point for the focus adjustment. Therefore, the change of the focus adjustment due to the manual operation need not be reversed or registered.

The driving connection may comprise electronic, electrical and / or mechanical connecting elements.

A simple embodiment can provide that a mechanical or electrical actuating means for disengaging and engaging, so separating and closing, the driving connection is present.

The invention will now be described in more detail with reference to an embodiment, but is not limited to this embodiment. Further embodiments result from combining one or more features of the claims with each other and / or with one or more features of the embodiment.

• 1 eine erfindungsgemäße Wärmebildkamera in einer Schrägansicht von vorn,
• 2 die Wärmebildkamera gemäß 1 in einer Schrägansicht von hinten,
• 3 eine Prinzipskizze zur Erläuterung des erfindungsgemäßen Verfahrens und
• 4 ein stark vereinfachtes Ablaufdiagramm eines erfindungsgemäßen Verfahrens.

It shows

• 1 a thermal imaging camera according to the invention in an oblique view from the front,
• 2 the thermal imaging camera according to 1 in an oblique view from behind,
• 3 a schematic diagram for explaining the method according to the invention and
• 4 a highly simplified flow diagram of a method according to the invention.

1 and 2 show different oblique views of a whole designated 1 thermal imaging camera.

Inside the thermal imager 1 is an IR sensor arrangement, not shown 2 , For example, a FPA, arranged and for receiving IR images through an IR lens 3 set up.

The thermal imager 1 thus has a recording device 28 which at least the IR sensor arrangement 2 and the IR lens 3 includes.

The IR lens 3 or the receiving device 28 has a trained in a conventional manner focusing 7 to change the focus setting and thus the distance setting of the IR lens 3 on.

The thermal imager 1 is as a handheld device with a handle 4 formed, being in the range of the handle 4 a trigger element 5 is arranged, with which the recording of an IR image can be triggered. The recording of IR images can also be triggered automatically, for example, at predetermined or fixed time intervals.

The recorded IR images are - optionally after a picture processing - on a display means 6 , For example, a display or the like, displayed.

The display means 6 and / or the operating mode of the thermal imager 1 can by actuators 8th to be influenced.

For example, can be set to the thermal imaging camera 1 automatically takes a sequence of IR images.

3 shows a sequence of captured IR images 9 . 10 . 11 . 12 . 13 , which according to the inventive method for different focus settings 14 . 15 . 16 . 17 . 18 of the IR lens 3 were recorded. In the thermal imager 1 is now an image processing unit 19 designed and set up with which to each of the recorded IR images 9 . 10 . 11 . 12 . 13 a measure can be calculated. This measure 20 In this case, it depends on the proportion of high frequencies in a transformed image in a monotonously increasing manner and thus makes it possible to make a statement about the number of edges in the respective IR image 9 . 10 . 11 . 12 . 13 , The measure 20 can also directly from the number of edges in the IR image 9 . 10 . 11 . 12 . 13 depend.

For example, the transformed images may be Fourier transform or other suitable base transformation in the image processing unit 19 be calculated. In this case, the measure 20 as a sharpening function S (p) from the mean over the amounts of the calculated with a film transformation spectrum. For this purpose, the magnitude, ie the length of the connection vector between imaginary and real part of the signal of the transformed image, can be calculated and summed over the entire transformed image.

To take the IR pictures 9 . 10 . 11 . 12 . 13 becomes the focus adjustment 21 of the IR lens 3 with a step size 29 varies, which is chosen constant in the embodiment and which in further embodiments variable to the course of the associated measure 20 can be adjusted. The automated image recording method of the invention now provides that first of all an initial focus adjustment 14 starting with a monotone ascending sequence of focus settings 14 . 15 . 16 . 17 . 18 is passed through, for each of which an associated IR image 9 . 10 . 11 . 12 . 13 is recorded.

In this way, support points for the function of the measure result 20 as function S (p) of focus adjustment 21 , This measure 20 is in a correspondingly equipped evaluation unit 29 calculated, which together with or separately from the image processing unit 19 can be trained.

Subsequently, an interpolation curve 23 calculated for the determined support values.

For this interpolation curve 23 is then an optimal value by means known per se 24 , in the present example a maximum, determined.

This optimal value 24 thus separates a monotonously rising section 25 the measure S (p) from a monotone falling section 26 ,

For example, this optimal value 24 also be determined at least approximately, by first a monotonically increasing sequence of focus settings 14 . 15 . 16 . 17 is passed through for these focus settings 14 . 15 . 16 . 17 each IR images 9 . 10 . 11 . 12 recorded and evaluated as described. It can be seen that the associated numerical measures initially form a sequence of numbers with predetermined, here increasing monotony. Since that to the focus adjustment 17 appropriate measure 20 opposite to the measure 20 to the focus adjustment 16 has fallen while the measures for the sequence of focus settings 14 . 15 . 16 have increased monotonically, the numerical order of the measures thus deviates when reaching the focus setting 17 from the given monotony and now has a falling monotony. Therefore, the focus setting now becomes more refined 21 Reduced again and IR images are taken again and evaluated to the point of optimal value 24 the measure 20 to determine more precisely.

At the end of the procedure becomes the focus adjustment 27 on the IR lens 3 set which to the optimum value 24 the measure 20 corresponds.

This is in the thermal imaging camera 1 a not further apparent control unit 30 formed, which the focusing unit 7 correspondingly activates via a control connection, not shown.

Should the focusing unit 7 be operated manually, the drive connection can be disconnected. If the automatic focus adjustment is to be retrieved by the method according to the invention, then the drive connection only needs to be closed. Resetting the adjusted during manual operation of the focus adjustment to a defined starting position is not necessary because the inventive method can start from any initial situation.

As by the schematic representation of the recorded IR images 9 . 10 . 11 . 12 . 13 indicates that points to the optimum focus setting 27 proper IR image 11 the largest number of edges and thus the best focus on. For the rest, "blurry" IR images 9 . 10 . 12 . 13 On the other hand, the edges of the grid pattern used by way of example are smeared or no longer depicted.

4 shows a flowchart of a method according to the invention in a highly simplified representation.

In this method, a status flag is accessed which can assume at least the values "Init", "FineTune", "Autofocus" and "Exit". Here, "Init" denotes a traversing section that is executed at the beginning of the method according to the invention. Autofocus refers to one Normal operating mode and "Exit" indicates the successful completion of the method according to the invention.

The value "FineTune" of the status flag designates an option with which the result of the method according to the invention can be further improved and which will be described in more detail below.

First, in a startup step 31 the method according to the invention is started, wherein the status flag is initially set to the value "Init". However, it can also be provided that the value "Autofocus" is set directly.

Subsequently, in a status query 32 queried the status flag. If the value equals "Init", "Autofocus" or "FineTune" then the branch will be 33 pursued.

On the other hand, if the value of the status flag is equal to "Exit", the branch becomes 34 followed up, and the process ends in step 44 ,

In the branch 33 becomes a focus setting 14 . 15 . 16 . 17 . 18 set and it gets an IR image 9 . 10 . 11 . 12 . 13 added. Here, a focus setting is selected, which deviates in a predetermined direction, ie monotony, by a predetermined increment from a previous focus setting.

The recorded IR image 9 . 10 . 11 . 12 . 13 is in an image editing step 36 transformed into a transformed image with a base transformation for a given image area.

In an image evaluation step 37 becomes at least one measure sVal from the transformed image 20 (see. 3 ) derived.

This measure sVal - or 20 depends monotonically on the proportion of high frequencies in the transformed image and / or on the number of edges in the IR image 9 . 10 . 11 . 12 . 13 from.

Now in a renewed status query 38 the status flag is read out again, and it becomes for the values "Autofocus" and "FineTune" the monotone evaluation 39 while for the case "Init" directly the evaluation section 14 is jumped.

In the monotone evaluation 39 becomes the derived measure sVal 20 is compared with a measure sVal 'derived in the previous cycle. If the new, derived measure sVal is greater than the previously derived measure sVal ', then the branch 41 followed up, otherwise it will cross the branch 42 the evaluation section 40 started.

In the evaluation section 40 is checked for the value "Init" of the status flag, whether the previously recorded sequence of measures sVal 20 detects a maximum or exceeds a threshold. If this is the case, then the direction or the sign, with which in the image evaluation step 37 the focus setting is changed, vice versa. This inversion changes the monotony of the sequence of approached focus settings.

If this is not the case, then the predetermined direction of the change of the focus settings is retained.

In the evaluation section 40 is checked for the value "Autofocus", whether in the image acquisition step 35 The direction used to change the focus settings has already been reversed. If this is the case, the status flag is set to the value "Exit". Otherwise, in the image pickup step 35 the direction used to change the focus settings, ie the monotony for the sequence of focus settings, vice versa, and this is noted in a corresponding memory area.

Became the branch 41 has passed, so no maximum in the sequence of derived measures sval 20 detected and the monotony of the sequence of measures has not changed yet.

It is therefore in the step 43 the value "autofocus" of the status flag is retained and the value of the newly derived metric sVal is written in place of the metric sVal 'derived in the previous cycle step.

After processing the evaluation section 40 or the step 43 the procedure jumps back to the status query 32 , and a new cycle is executed.

By registering and monitoring the number of reversals, that is to say the number of monotony changes introduced by the method in the sequence of focus settings, a usable termination condition can be formulated for the method according to the invention, according to which at least approximately optimal focus adjustment is present.

Upon reaching the final step 44 is thus on the IR lens 3 the focus setting 27 to the optimum value 24 the measure 20 set.

The result can be further refined by setting a value "FineTune" in addition to the "Autofocus" value of the status flag, in which the routine described above for "Autofocus" uses a refined step size and / or several times in succession to determine a mean value for the optimal measure or the corresponding focus setting can be determined.

In this way it is even possible to eliminate hysteresis effects of the motor control and other effects as far as possible.

With the thermal imager 1 For example, an image acquisition method is set up for which a sequence of focus settings 14 . 15 . 16 . 17 . 18 an IR lens 3 the thermal imager 1 IR images 9 . 10 . 11 . 12 . 13 be recorded, taking the captured IR images 9 . 10 . 11 . 12 . 13 at least in regions with a base transformation converted into transformed images, wherein from the transformed images in each case a measure 20 derived and the associated focus setting 14 . 15 . 16 . 17 . 18 is assigned, the measure 20 monotone of the high frequency portion in the transformed image and / or edges in the IR image 9 . 10 . 11 . 12 . 13 depends, and being derived from the derived measures 20 an optimal focus setting 27 determined and on the IR lens 3 is set.

Claims (10)

Image recording method for IR images (9, 10, 11, 12, 13), wherein for different focus settings (14, 15, 16, 17, 18) IR images (9, 10, 11, 12, 13) are recorded, wherein from the recorded IR images (9, 10, 11, 12, 13) for a given image area with a base transformation transformed images are calculated, wherein derived from the transformed images in each case a measure (20) and the associated focus setting (14, 15, 16, 17, 18), the measure (20) being monotonically dependent on the high frequency portion in the transformed image and on a number of edges in the IR image (9, 10, 11, 12, 13) from the derived and associated measures (20) an optimal measure (24) is determined, wherein one of the optimal measure (24) corresponding focus setting (27) is set and wherein when determining the optimum measure (24) an interpolation curve (23) the derived measures (20) as a function of zugeo rdneten focus settings (21) is calculated. Image recording method after Claim 1 , characterized in that the focus adjustment (21) is changed with a monotonous sequence of distance settings until the sequence of associated measures (20) has an extreme value (24). Image recording method after Claim 1 or 2 , characterized in that the focus adjustment (21) is changed in a first step with a monotone sequence of distance settings with a first monotony and is changed in a second step with a monotonous sequence of distance adjustments with opposite monotony to the first monotony as soon as the sequence the assigned measures (20) in the first step of a predetermined monotony (25, 26) differs. Image recording method according to one of Claims 1 to 3 , characterized in that the focus adjustment (21) with a step size (29) is changed, which is determined by the direction and / or the strength of the change of the associated measures (20). Image recording method according to one of Claims 1 to 4 , characterized in that in the derivation of the measure (20) from a transformed image, an average of the magnitudes of the transformed image is formed in a frequency range and / or that the captured IR images (9, 10, 11, 12, 13) and / or the transformed images are smoothed before deriving the measure (20). Image recording method according to one of Claims 1 to 5 , characterized in that a predetermined focus adjustment (27) is selected if the sequence of the derived and associated measures (20) does not allow the determination of an optimal measure (24). Image recording method according to one of Claims 1 to 6 , characterized in that the transformed images are calculated with a 2D Fourier transformation, in particular with an FFT or with a DFT, and / or with a base transformation with wavelets and / or with a cosine transformation. Image recording method according to one of Claims 1 to 7 , characterized in that the focus settings (21) are each designated by a distance setting of a lens (3), wherein the IR images (9, 10, 11, 12, 13) are received by the lens (3), and / or that a drive connection for driving the focus settings is disengaged, after which a focus adjustment can be manually set or set. Thermal imaging camera (1) with a recording device (28) for IR images (9, 10, 11, 12, 13), wherein the receiving device (28) has an adjustable Focusing unit (7) has, characterized in that an image processing unit (19) is formed, which is adapted to calculate a base transformation of a predetermined image area of a captured IR image (9, 10, 11, 12, 13) in a transformed image that an evaluation unit (21) is provided, which is set up to calculate a measure (20) for the respective transformed image, wherein the measure (20) is monotone of the proportion of high frequencies in the respective transformed image and of a number of edges in the IR Image (9, 10, 11, 12, 13) that a control unit (30) is provided, with which the focusing unit (7) can be controlled in such a way that the resulting IR image (9, 10, 11, 12, 13) assumes an optimum value (24), and that the thermal imaging camera for carrying out a method according to one of the Claims 1 to 8th is set up. Thermal Imaging Camera (1) after Claim 9 characterized in that a driving connection between the focusing unit (7) and the control unit (30) is disengageable and disengageable, wherein with disengaged drive connection the focusing unit (7) is manually operable, and / or that the driving connection electronic, electrical and / or Includes mechanical fasteners and / or that a mechanical or electrical actuating means for disengaging and engaging, so disconnecting and closing, the driving connection is present.

Images