JP5279236B2 - Target imaging detector - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a target imaging detector that achieves reduction in misdetection of a target. <P>SOLUTION: The target imaging detector includes first and second waveband imaging parts 1 and 2 for imaging the same target in two different wavebands; an error function calculating part 3 for substituting an image output signal y, from the first waveband imaging part into a function containing an error function expressed by f(y)=(Aexp(-By))/(1+erf(C+Dy)), by using parameters A-D to output the calculation results;an exponential function calculation part 4 for substituting an image output signal x from the second waveband imaging part into an exponential function, expressed by g(x)=exp(Ex<SP>2</SP>), by using a parameter E to output calculation results; and a target determining part 5 for determining that a detected signal is a signal from the target, when the calculation results output from the error function calculating part and the exponential function calculating part satisfy the relation g(x)&le;f(y). <P>COPYRIGHT: (C)2009,JPO&amp;INPIT

Description

  The present invention relates to a target imaging detection apparatus that automatically detects a target by imaging the sensor, and particularly relates to signal processing when an infrared sensor is used.

  Among sensors used for automatically detecting a target, infrared sensors that can be used day and night are widely used. However, since infrared rays are always emitted from all objects having temperature, the target signal detected by the infrared sensor is affected by thermal radiation from the surroundings of the target. In addition, if sunlight, scattered light in the atmosphere, or a radiation spectrum from around the target overlaps with the target radiation spectrum, it is difficult to separate both radiation spectra by single-wavelength processing, and thus misdetect the target. At this time, the target that causes false detection with the target is called clutter.

  Therefore, two-wavelength processing using the fact that the infrared radiation spectrum differs depending on the type of object, temperature, emissivity, etc. is used for target detection. In this process, signal detection is performed in different wavelength bands, and only the target signal is detected from the difference in radiation spectrum characteristics between the target and the clutter. Here, a threshold for identifying the target signal and the clutter signal is required.

  For example, in Patent Document 1 below, an image output signal captured in two different wavelength bands is converted into coordinates in an orthogonal coordinate system with each wavelength band as an axis, and then an elliptical shape created in advance on this coordinate system The threshold filter is used to determine whether the output signal is the target signal.

Japanese Patent Laid-Open No. 08-068844

  When light propagates through the atmosphere, it is affected by absorption and scattering by molecules present in the atmosphere, and the intensity of light attenuates exponentially as the propagation distance increases. Due to such atmospheric loss of light, the target signal intensity detected by the sensor changes nonlinearly with the distance between the target and the sensor, and the target signal when the distance between the target and the sensor changes with a uniform probability. The distribution does not follow a normal distribution. As a result, the target signal distribution when the image output signal captured in different wavelength bands is converted into the coordinates of the orthogonal coordinate system with each wavelength band as an axis is not an elliptical shape, and the equal probability density curve of the target signal distribution The egg shape as shown in FIG. Therefore, by setting the threshold for determining the target as the equiprobability density curve shown in FIG. 3, the threshold filter area can be minimized while maintaining high detection performance, and the false detection of background clutter can be minimized. Is possible.

  An object of the present invention is to realize a reduction in target false detection by setting a target determination threshold as described above in a target imaging detection apparatus including a two-wavelength infrared sensor that performs automatic target detection.

This invention uses parameters A to D for the first wavelength band imaging unit and the second wavelength band imaging unit for imaging the same target in two different wavelength bands, and the image output signal y from the first wavelength band imaging unit. F (y) = (Aexp (-By)) / (1 + erf (C + Dy)) (B / (2CD) ≧ 1) Substitute the error function calculation unit to output and the image output signal x from the second wavelength band imaging unit into the exponential function represented by g (x) = exp (Ex ² ) using the parameter E, and the calculation result Determines that the detection signal is the target signal when the exponential function calculation unit to be output and the calculation results output from the error function calculation unit and the exponential function calculation unit have a relationship of g (x) ≦ f (y) And a target determination processing unit.

  According to the present invention, it is possible to reduce target false detection in a target imaging detection apparatus including a two-wavelength infrared sensor that performs automatic target detection.

Embodiment 1 FIG.
FIG. 1 is a diagram showing a configuration of a target imaging detection apparatus according to Embodiment 1 of the present invention. The target imaging detection device includes a first wavelength band imaging unit (infrared sensor) 1, a second wavelength band imaging unit (infrared sensor) 2, an error function calculation unit 3, an exponential function calculation unit 4, and a target determination processing unit 5. Has been. The same target is imaged in different wavelength bands of the first wavelength band imaging unit 1 and the second wavelength band imaging unit 2. At this time, the wavelength band of the first wavelength band imaging unit 1 selects a wavelength band from which a large signal output can be obtained from the target radiation, and the wavelength band of the second wavelength band imaging unit 2 outputs a large signal from the radiation of the background clutter. Is selected. Then, the image output signal from the first wavelength band imaging unit 1 is calculated by substituting the error output calculation unit 3 into a function including an error function, and the image output signal from the second wavelength band imaging unit 2 is exponentially calculated. In part 4, the calculation is performed by substituting it into the exponential function, and the calculation result is output. Finally, the target determination processing unit 5 determines whether the detected signal is the target signal based on the calculation results of the error function calculation unit 3 and the exponent function calculation unit 4.

  FIG. 2 is a signal distribution when image output signals obtained by imaging the same target by the first wavelength band imaging unit 1 and the second wavelength band imaging unit 2 are expressed by coordinates in an orthogonal coordinate system with each imaging wavelength band as an axis. The result of having been simulated by the computer is shown. In FIG. 2, the output of the first wavelength band imaging unit 1 is shown on the vertical axis (y axis), and the output of the second wavelength band imaging unit 2 is shown on the horizontal axis (x axis). FIG. 2 shows a simulation result when parameters such as the target-sensor distance, sensor temperature, background temperature, and background reflectance are changed with a uniform probability.

  At this time, since the target signal detected by the sensor is affected by absorption and scattering by molecules in the atmosphere, the target signal intensity attenuates exponentially as the distance between the target and the sensor increases. Thus, since the target signal intensity changes nonlinearly depending on the distance between the target and the sensor, the target signal distribution from the first wavelength band imaging unit 1 that selects the wavelength band in which the characteristics of the target radiation spectrum intensity appear (FIG. 2). In particular, the distribution of signals appearing in the vertical axis direction does not follow the normal distribution.

  On the other hand, the characteristics of the target radiation spectrum intensity do not appear so much in the image output signal when the target is imaged by the second wavelength band imaging unit 2 that selects the wavelength band in which the characteristics of the radiation spectrum intensity of the background clutter appear. The sensor output is dominant in the image output signal from the wavelength band imaging unit 2, and the target signal distribution at this time (the distribution of signals appearing in the horizontal axis direction in FIG. 2) follows a normal distribution.

  Therefore, the target signal area on the orthogonal coordinate system with the respective wavelength bands of the first wavelength band imaging unit 1 and the second wavelength band imaging unit 2 as an axis is a signal distribution due to sensor noise corresponding to the signal intensity in each axis direction. When considering the spread, it becomes as follows.

In Equation 1, N (I) is the target signal distribution in the output (y-axis) direction of the first wavelength band imaging unit 1, and I ₀ is the wavelength band axis of the first wavelength band imaging unit 1 determined by the target detection distance ( (y-axis) direction target signal threshold, σ represents a signal distribution spread coefficient due to sensor noise corresponding to each axis direction.

  Accordingly, the target signal area is as follows using the parameters A to E.

exp (Ex ² ) ≦ (Aexp (−By)) / (1 + erf (C + Dy)) Equation 2

  Here, the following relational expression holds for each parameter of Expression 2.

    B / (2CD) ≧ 1 Formula 3

  FIG. 3 shows an example of the equiprobability density curve of the target signal area expressed by Equation 2, and indicates that the white area (that is, the inner egg type) has a higher distribution density. As in the example shown in FIG. 3, the target signal area has an egg shape instead of an elliptical shape. This can also be confirmed from the target signal region simulation results shown in FIG.

  Therefore, in the present invention, as shown in FIG. 1, an error function calculation unit 3 is arranged downstream of the first wavelength band imaging unit 1, and an exponential function calculation unit 4 is arranged downstream of the second wavelength band imaging unit 2, so that each calculation unit It is determined whether the detected signal is a target from the output result of the arithmetic processing described later in 3 and 4.

  In the error function calculation unit 3 in FIG. 1, the image output signal y from the first wavelength band imaging unit 1 is substituted into a function including an error function represented by the following equation, and the calculation result is output.

    f (y) = (Aexp (−By)) / (1 + erf (C + Dy)) Equation 4

  1 substitutes the image output signal x from the second wavelength band imaging unit 2 for the exponential function shown in the following equation and outputs the calculation result.

g (x) = exp (Ex ² ) Formula 5

  Then, the target determination processing unit 5 determines that the detected signal is the target signal when each output signal from the error function calculation unit 3 and the exponent function calculation unit 4 satisfies the following expression.

    g (x) ≦ f (y) Equation 6

  FIG. 4 shows a conceptual diagram for detecting the target signal by the signal processing in the simulation result of the target signal region shown in FIG. Here, FIG. 4 shows an example of a computer simulation result including the background clutter signal distribution in addition to the target signal distribution of FIG. 2, and the target signal determination threshold of Expression 6 for determining the target signal is indicated by a broken line. . As shown in FIG. 4, the target signal determination threshold in the target imaging detection apparatus of the present invention is set along the target signal region, and the target signal and the background clutter signal can be identified.

In this embodiment, not only the shape as the target signal determination threshold shown in FIG. 4 but also the target signal determination threshold (target signal region) having the shape as shown by the broken line in FIG. It is possible to simplify the arithmetic processing of the apparatus without lowering. In the example shown in FIG. 5, when the value of y when the equation 2 x is the maximum value x _max and y _max, y ≦ y for the _max of the output signal of the formula 5 g (x) ≦ f ( The target signal determination process is performed using y). On the other hand, target signal determination processing is performed using | x | ≦ x _max for an output signal of y> y _max . This is a setting of a target signal determination threshold value corresponding to the target signal region when it is assumed that the output signal when the distance between the target and the sensor approaches as much as possible becomes extremely large. At this time, in the example shown in FIG. 5, the target signal determination threshold value in the output signal y> y _max is shown by a straight line, but it may be a curve along the signal area of the background clutter assumed when performing the target detection. .

  As described above, according to the present invention, since the shape of the threshold value for determining whether or not the detected signal is the target signal matches the target signal region, it is possible to reduce erroneous detection of the target signal and the background clutter signal. Yes, the reliability of the target imaging detection device can be improved.

Embodiment 2. FIG.
In the first embodiment, the error function calculation unit 3 and the exponent function calculation unit 4 perform calculation using the same target image output signals output from the first wavelength band imaging unit 1 and the second wavelength band imaging unit 2. Is going. Here, the image output signals from the first wavelength band imaging unit 1 and the second wavelength band imaging unit 2 include not only the signal from the target but also the signal from the background. It is easy to be affected, and the calculation accuracy in the error function calculation unit 3 and the exponent function calculation unit 4 and the accuracy of the target determination processing in the target determination processing unit 5 are lowered.

  Therefore, an embodiment for removing the influence of signal level fluctuations due to background changes on target determination processing accuracy will be described.

  FIG. 6 shows a target imaging detection apparatus according to the first embodiment, in which a function for removing the influence of signal level fluctuations on the target determination processing accuracy due to changes in the background and sensor light reception sensitivity is added. It is a figure which shows the structure of a target imaging detection apparatus, and the code | symbol 1-5 is the same as that of Embodiment 1. FIG. Here, 6 is a first wavelength band high-pass filter, and 7 is a second wavelength band high-pass filter.

  In the second embodiment, the first wavelength band high-pass filter 6 is disposed downstream of the first wavelength band imaging unit 1, and the second wavelength band high-pass filter 7 is disposed downstream of the second wavelength band imaging unit 2. The high-pass filter corresponding to each imaging unit 1, 2 calculates the difference between the image output signal obtained from each imaging unit and the average signal intensity of each image, and removes the offset component due to the background signal included in each image output signal To do. As a result, only the target signal is used for the subsequent signal processing, so that it is possible to remove the influence of the signal level fluctuation caused by the background change on the target determination processing accuracy.

  In this embodiment, by extracting only the high-frequency component of the image output signal obtained from each imaging unit by the high-pass filter corresponding to each imaging unit, the luminance change of each image output signal is compared with the surrounding pixels. Only large pixels, that is, pixels for which a target signal is detected can be extracted. As a result, the target detection performance of the apparatus can be improved.

  From the above, the target imaging detection device according to this embodiment has a high-pass filter disposed after each imaging unit to remove the influence of signal level fluctuations due to background changes on target determination processing accuracy. It is possible to realize a target imaging detection device whose performance does not deteriorate in use conditions. Also, by extracting only the high-frequency component of the image output signal obtained from each imaging unit, it becomes easier to detect the target signal, and the target detection performance of the apparatus can be improved.

Embodiment 3 FIG.
In the first and second embodiments, the wavelength band of the first wavelength band imaging unit 1 selects a wavelength band that provides a signal output with a large target radiation, and the wavelength band of the second wavelength band imaging unit 2 includes background clutter. A wavelength band is selected in which a signal output with large radiation is obtained. At this time, if the wavelength bands of the first wavelength band imaging unit 1 and the second wavelength band imaging unit 2 are not selected as described above (in the case of FIG. 2), the target signal region does not follow Equation 2. As a result, the calculation process performed by the error function calculation unit 3 and the exponent function calculation unit 4 is not performed correctly, and the accuracy of the target determination process performed by the target determination process unit 5 is reduced.

  Therefore, an embodiment in which a sensor having an arbitrary imaging wavelength band can be applied to the first wavelength band imaging unit 1 and the second wavelength band imaging unit 2 will be described.

  FIG. 7 shows a target obtained by selecting a wavelength band in which the wavelength bands of the first wavelength band imaging unit 1 and the second wavelength band imaging unit 2 are different from the above (in the case of FIG. 2) in the target imaging detection apparatus of the first embodiment. It is a figure which shows the structure of the target imaging detection apparatus of this Embodiment 3 which added the function for preventing the fall of a determination processing precision, and the code | symbol 1-5 is the same as that of Embodiment 1. FIG. Here, 8 is a first wavelength band correction unit, and 9 is a second wavelength band correction unit.

  In the third embodiment, the first wavelength band correction unit 8 is arranged downstream of the first wavelength band imaging unit 1, and the second wavelength band correction unit 9 is arranged downstream of the second wavelength band imaging unit 2.

  FIG. 8A is a conceptual diagram of the target signal region when the wavelength bands of the first wavelength band imaging unit 1 and the second wavelength band imaging unit 2 are selected so as to have the above-described relationship. FIG. 8B shows a conceptual diagram of the target signal area when the wavelength band is selected. FIG. 8 shows an orthogonal coordinate system with the wavelength bands of the imaging units 1 and 2 as axes as in FIG. 2, and a region surrounded by a dotted line shows a target signal region. 8A and 8B have different wavelength bands for detecting the target radiation, the image output signals obtained from the respective imaging units differ between FIGS. 8A and 8B, and the target signal region. Are different depending on (a) and (b) of FIG.

  At this time, the above expression 2 is an expression representing the target signal area in the case of FIG. 8A, and cannot be applied to the target signal area as shown in FIG. 8B. Therefore, coordinate transformation is performed so that Equation 2 can be applied to the target signal region shown in FIG.

  One of the methods for performing coordinate conversion of the target signal region is a method for performing principal component analysis. In this method, a principal component analysis of the target signal region in FIG. 8B is performed, an orthogonal coordinate system having the principal component and a component orthogonal to the principal component as axes is obtained, and coordinate transformation is performed so that Equation 2 can be applied. It is a method to do. At this time, the principal component direction indicates a signal distribution direction (a signal distribution appearing in the vertical axis direction in FIG. 2) in which a large signal can be detected from the target radiation. A signal distribution direction in which only a small signal can be detected, that is, a signal distribution direction (a signal distribution appearing in the horizontal axis direction in FIG. 2) in which a signal output with large radiation of the background clutter is obtained is shown. Then, the coordinate axes in FIG. 8B are rotated so that the orthogonal coordinate system obtained as a result of the principal component analysis matches the orthogonal coordinate system in FIG. Accordingly, the target signal region in FIG. 8B can be handled in the same manner as the target signal region in FIG. 8A, and Expression 2 can be applied.

  In this embodiment, based on the coordinate axis rotation data obtained by the principal component analysis, the first wavelength band correction unit 8 that performs the coordinate conversion calculation of the image output signal from the first wavelength band imaging unit 1, the second wavelength band A second wavelength band correction unit 9 that performs a coordinate conversion calculation of an image output signal from the imaging unit 2 is disposed in the subsequent stage of each imaging unit. Then, by performing signal processing on the output signal from each correction unit in the same manner as in the first or second embodiment, it is possible to prevent a decrease in target determination processing accuracy.

  From the above, the target imaging detection device of this embodiment can be used without being affected by the imaging wavelength bands of the first wavelength band imaging unit 1 and the second wavelength band imaging unit 2.

  Needless to say, the present invention includes possible combinations of the above embodiments.

It is a figure which shows the structure of the target imaging detection apparatus by Embodiment 1 of this invention. A signal distribution state when an image output signal obtained by imaging the same target by the first wavelength band imaging unit and the second wavelength band imaging unit according to the present invention is expressed by coordinates of an orthogonal coordinate system having each imaging wavelength band as an axis. It is the figure which showed the result of having simulated by. It is the figure which showed the example of the equal probability density curve of the target signal area | region in this invention. FIG. 3 is a diagram illustrating an example of a computer simulation result including a background clutter signal distribution in addition to the target signal distribution of FIG. 2. FIG. 6 is a diagram for explaining a case where a target signal determination threshold (target signal region) different from that in FIG. 5 is set. It is a figure which shows the structure of the target imaging detection apparatus by Embodiment 2 of this invention. It is a figure which shows the structure of the target imaging detection apparatus by Embodiment 3 of this invention. It is a conceptual diagram of the coordinate for demonstrating the coordinate transformation of the target signal area | region in Embodiment 3 of this invention.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 1st wavelength band imaging part, 2 2nd wavelength band imaging part, 3 Error function calculating part, 4 Exponential function calculating part, 5 Target determination processing part, 6 1st wavelength band high pass filter, 7 2nd wavelength band high pass filter, 8 First wavelength band correction unit, 9 Second wavelength band correction unit.

Claims (4)

A first wavelength band imaging unit and a second wavelength band imaging unit for imaging the same target in two different wavelength bands;
F (y) = (Aexp (-By)) / (1 + erf (C + Dy)) (however, B / (2CD ) ≧ 1) An error function operation unit that outputs an operation result by substituting it into a function including an error function represented by
An exponential function computing unit that outputs an operation result by substituting the image output signal x from the second wavelength band imaging unit into an exponential function represented by g (x) = exp (Ex ² ) using the parameter E;
A target determination processing unit that determines that the detection signal is a target signal when the calculation results output from the error function calculation unit and the exponential function calculation unit have a relationship of g (x) ≦ f (y);
A target imaging detection apparatus comprising: The first wavelength band imaging unit images the target in a wavelength band where a large signal output can be obtained from the target radiation, and the second wavelength band imaging unit images the target in a wavelength band where a large signal output can be obtained from the background clutter radiation. The target imaging detection apparatus according to claim 1, wherein:   A first wavelength band high-pass filter and a second wavelength band high-pass filter for removing the influence of signal level fluctuation due to a background change are provided at the subsequent stage of the first wavelength band imaging unit and the second wavelength band imaging unit. The target imaging detection device according to claim 1 or 2. A first wavelength band correction unit that performs coordinate axis conversion to an orthogonal coordinate system with the main component of each image output signal of the first wavelength band imaging unit and the second wavelength band imaging unit and a component orthogonal to the main component as axes; and The target imaging detection apparatus according to claim 1, further comprising a second wavelength band correction unit.

Images