A method and device for detecting fixed pattern noise of a camera module

Abstract

This invention discloses a method for detecting fixed pattern noise in a camera module, comprising the following steps: Step 100: obtaining a test image of the camera module; Step 200: decomposing the test image into multiple single-channel images according to the number of channels in the test image; Step 300: calculating the single-row pixel value and single-column pixel value of all single-channel images; Step 400: determining whether fixed pattern noise exists in the test image based on the single-row pixel value and single-column pixel value of all single-channel images. This fixed pattern noise detection method can improve the detection and recognition accuracy of fixed pattern noise. This invention also discloses a fixed pattern noise detection device for a camera module, used to implement the above method.

Description

Technical Field

[0001] This invention relates to the field of camera module testing, and more particularly to a method and apparatus for detecting fixed pattern noise in camera modules. Background Technology

[0002] A camera module is a modular component that integrates an image sensor, optical lens, image processing chip, and other related circuitry. This modular design simplifies camera integration, allowing it to be easily integrated into various devices such as smartphones, tablets, smart home devices, and industrial vision systems.

[0003] Fixed pattern noise (FPN) is noise that appears at a fixed location in an image. This noise repeats in the same position in multiple images captured or photographed. Its causes include uneven response rates of the image sensor, optical lens defects, and manufacturing errors of the image sensor. Therefore, camera modules undergo fixed pattern noise testing before leaving the factory to screen out defective products and prevent them from reaching end customers.

[0004] For example, Chinese Patent Application No. CN201580052995.9 discloses a fixed pattern noise calculation method based on an infrared camera device. The infrared camera device has: an imaging element comprising a plurality of infrared detection pixels arranged in two dimensions; and an aperture disposed closer to the subject than the imaging element. The fixed pattern noise calculation method includes: a fixed pattern noise calculation step, when capturing a dynamic image, obtaining first image data captured by the imaging element with the F-value of the aperture as a first value and second image data captured by the imaging element with the F-value of the aperture as a second value, and calculating the fixed pattern noise contained in the image data captured by the imaging element based on the obtained first image data, second image data, first value, and second value.

[0005] However, image sensors have multiple channels, each of which is only responsible for recording the brightness information of the corresponding color. When there is a color shift problem in the image sensor, the channel corresponding to the color shift will have a problem of higher brightness, which will lead to a large difference in brightness between this channel and other channels, and thus be misjudged as noise.

[0006] In addition, existing fixed pattern noise testing methods can only detect and identify stripe noise that appears in a whole row or column. When the fixed pattern noise does not appear in a whole row or column of stripes, it cannot be detected and identified. Summary of the Invention

[0007] To address the shortcomings of the prior art, the present invention provides a method and apparatus for detecting fixed pattern noise in a camera module, which can improve the detection and recognition accuracy of fixed pattern noise.

[0008] The technical problem to be solved by the present invention is achieved through the following technical solution:

[0009] A method for detecting fixed pattern noise in a camera module includes the following steps:

[0010] Step 100: Obtain test images of the camera module;

[0011] Step 200: Decompose the test image into multiple single-channel images according to the number of channels in the test image;

[0012] Step 300: Calculate the single-row pixel values ​​and single-column pixel values ​​for all single-channel images;

[0013] Step 400: Determine whether the test image contains fixed pattern noise based on the single row pixel value and single column pixel value of all single channel images.

[0014] Furthermore, in step 400, the step of determining whether the test image contains fixed pattern noise based on the single-row pixel values ​​and single-column pixel values ​​of all single-channel images is as follows:

[0015] Step 410: Group all single-row pixel values ​​and single-column pixel values ​​of each single-channel image to obtain multiple row data groups and column data groups. Each row data group has A adjacent single-row pixel values, and each column data group has A adjacent single-column pixel values, where A≥2.

[0016] Step 420: Calculate the average of all row data groups and column data groups to obtain multiple row group averages and column group averages;

[0017] Step 430: Calculate the difference between the average row grouping and the average column grouping of three adjacent rows, and obtain multiple row grouping differences and column grouping differences;

[0018] Step 440: Filter out the maximum value of all row grouping differences and column grouping differences to obtain the maximum difference of row grouping and column grouping;

[0019] Step 450: Repeat steps 410-440 until the maximum difference between row groups and column groups of all single-channel images is obtained;

[0020] Step 460: Calculate the row discrete value and column discrete value of the test image based on the maximum difference between row groups and the maximum difference between column groups of all single-channel images;

[0021] Step 480: Determine whether the test image contains fixed pattern noise based on the row discrete value, column discrete value and preset discrete threshold.

[0022] Furthermore, in step 480, when determining whether the test image has fixed pattern noise, if the row discrete value or column discrete value is greater than a preset discrete threshold, it is determined that the test image has fixed pattern noise; if both the row discrete value and the column discrete value are less than the preset discrete threshold, it is determined that the test image does not have fixed pattern noise.

[0023] Furthermore, between steps 460 and 480, the following steps are also included:

[0024] Step 470: Repeat steps 410-460 to obtain multiple row discrete values ​​and column discrete values ​​of the test image when the single-channel image is grouped with different A values.

[0025] Furthermore, in step 480, when determining whether the test image has fixed pattern noise, if one of the discrete values ​​of all rows or one of the discrete values ​​of all columns is greater than a preset discrete threshold, then the test image is determined to have fixed pattern noise; if all the discrete values ​​of all rows and columns are less than the preset discrete threshold, then the test image is determined not to have fixed pattern noise.

[0026] Furthermore, in step 400, the step of determining whether the test image contains fixed pattern noise based on the single-row pixel values ​​and single-column pixel values ​​of all single-channel images is as follows:

[0027] Step 410: Calculate the row discrete value and column discrete value of each single-channel image based on all single-row pixel values ​​and single-column pixel values;

[0028] Step 420: Determine whether there is fixed pattern noise in each single-channel image based on the row discrete value, column discrete value and preset discrete threshold;

[0029] Step 430: Repeat steps 410 and 420 to determine whether fixed pattern noise exists in all single-channel images;

[0030] Step 440: Based on the noise levels of all single-channel images, determine whether the test image contains fixed pattern noise.

[0031] Furthermore, in step 440, when determining whether the test image has fixed pattern noise, the test image is determined to have fixed pattern noise as long as one of the single-channel images is determined to have fixed pattern noise, and the test image is determined to be free of fixed pattern noise only when all single-channel images are determined to be free of fixed pattern noise.

[0032] Furthermore, the test images are obtained by the camera module in a dark environment.

[0033] Furthermore, during shooting, the brightness gain value of the camera module is increased.

[0034] A fixed pattern noise detection device for a camera module includes a processor and a memory electrically connected to the processor. The memory stores a computer program that the processor can execute. When the processor executes the computer program, it performs the aforementioned fixed pattern noise detection method.

[0035] The present invention has the following beneficial effects:

[0036] The fixed pattern noise detection method of the present invention decomposes the test image of the camera module into multiple single-channel images according to the number of its channels, and then calculates the single-row pixel value and single-column pixel value of each single-channel image separately to determine whether the test image has fixed pattern noise. When the brightness of a certain channel of the camera module is high due to color shift of the image sensor, since each single-channel image only has brightness information of one color, the calculation of single-row pixel value and single-column pixel value of the single-channel images corresponding to other colors is not affected by the single-channel images corresponding to the color shift color, thereby improving the detection and recognition accuracy of fixed pattern noise.

[0037] The fixed pattern noise detection method of the present invention determines whether there is fixed pattern noise in all single-channel images based on the dispersion of column pixel values ​​and row pixel values ​​of all single-channel images, rather than based on traditional edge detection values. When the fixed pattern noise does not appear as a whole row or a whole column of stripes, it can still be detected and identified based on the dispersion, thereby improving the detection and identification accuracy of fixed pattern noise. Attached Figure Description

[0038] Figure 1 This is a flowchart illustrating the steps of the fixed pattern noise detection method provided by the present invention.

[0039] Figure 2 This is a flowchart of step 400 in the fixed pattern noise detection method provided by the present invention.

[0040] Figure 3 This is a flowchart of step 400 in another fixed pattern noise detection method provided by the present invention.

[0041] Figure 4 This is a flowchart of step 400 in another fixed pattern noise detection method provided by the present invention. Detailed Implementation

[0042] The present invention will now be described in detail with reference to the accompanying drawings and embodiments, examples of which are shown in the drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0043] In the description of this invention, it should be understood that the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0044] Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first," "second," or "third" may explicitly or implicitly include one or more of that feature. In the description of this invention, "multiple" means two or more, unless otherwise explicitly specified.

[0045] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," "fixing," and "setting," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0046] Example 1

[0047] like Figure 1 As shown, a fixed pattern noise detection method for a camera module includes the following steps:

[0048] Step 100: Obtain test images of the camera module.

[0049] In step 100, the test image is captured by the camera module in a dark environment, which refers to an environment with very weak or almost no light. Since the non-uniformity of dark current is the main source of fixed pattern noise, in the dark environment, the image sensor of the camera module is only affected by the dark current, which can eliminate the noise influence of external light on the image sensor, thereby more accurately evaluating the dark signal non-uniformity of each pixel in the image sensor.

[0050] Specifically, during shooting, a black baffle can be placed in front of the optical lens of the camera module to block external light from entering the optical lens, thereby creating the dark field environment. Of course, the dark field environment is a commonly used testing environment for the camera module, and there are various methods to create such a dark field environment, which should not be limited to this embodiment.

[0051] Preferably, during shooting, the brightness gain value of the camera module is increased. By increasing the brightness gain value, the signal strength of the image sensor, including noise signals, can be improved, making the fixed pattern noise in the test image more obvious.

[0052] Ideally, the brightness gain value can be set to the maximum value that the camera module can support. Of course, the brightness gain value can also be set to an appropriate value according to the test requirements, as long as it can clearly identify the fixed pattern noise in the test image.

[0053] Preferably, the number of test images is multiple to eliminate the test error of a single image and improve the test accuracy.

[0054] Step 200: Decompose the test image into multiple single-channel images according to the number of channels of the test image.

[0055] In step 200, the image sensor of the camera module has multiple channels, each channel is only responsible for recording the brightness information of the corresponding color, and the single-channel image only has the brightness information recorded by one channel of the test image.

[0056] Taking the common Bayer array pixel structure as an example, each pixel array in the image sensor consists of four pixels: RGGB, namely one red pixel, one blue pixel, and two green pixels, which correspond to one red channel, one blue channel, and two green channels, respectively. That is, one pixel in the pixel array represents one channel, and even if there are multiple pixels of the same color in the pixel array, they represent different channels.

[0057] Step 300: Calculate the single-row pixel value and single-column pixel value of all single-channel images.

[0058] In step 300, all single-channel images are composed of m rows and n columns of pixels of corresponding colors, where m and n ≥ 2. For example, the red channel image is composed of m rows and n columns of red pixels, the blue channel image is composed of m rows and n columns of blue pixels, and the green channel image is composed of m rows and n columns of green pixels.

[0059] When calculating the pixel values ​​of a single row and a single column for each single-channel image, the brightness values ​​of all pixels in the first pixel row are summed and then divided by the number of pixels in the first pixel row to obtain the single-row pixel value R1. Then, the brightness values ​​of all pixels in the second pixel row are summed and then divided by the number of pixels in the second pixel row to obtain the single-row pixel value R2, and so on. The brightness values ​​of all pixels in the m-th pixel row are summed and then divided by the number of pixels in the m-th pixel row to obtain the single-row pixel value Rm. Next, the brightness values ​​of all pixels in the first pixel column are summed and then divided by the number of pixels in the first pixel column to obtain the single-column pixel value C1. The brightness values ​​of all pixels in the second pixel column are summed and then divided by the number of pixels in the second pixel column to obtain the single-column pixel value C2, and so on. The brightness values ​​of all pixels in the n-th pixel column are summed and then divided by the number of pixels in the n-th pixel column to obtain the single-column pixel value Cn.

[0060] By performing the above calculations sequentially on all single-channel images, the single-row pixel values ​​and single-column pixel values ​​of all single-channel images are obtained sequentially. Specifically, if the image sensor uses a Bayer array, the single-row pixel values ​​and single-column pixel values ​​of one red channel image, one blue channel image, and two green channel images need to be obtained sequentially.

[0061] Step 400: Determine whether the test image contains fixed pattern noise based on the single row pixel value and single column pixel value of all single channel images.

[0062] The fixed pattern noise detection method of the present invention decomposes the test image of the camera module into multiple single-channel images according to the number of its channels, and then calculates the single-row pixel value and single-column pixel value of each single-channel image separately to determine whether the test image has fixed pattern noise. When the brightness of a certain channel of the camera module is high due to color shift of the image sensor, since each single-channel image only has brightness information of one color, the calculation of single-row pixel value and single-column pixel value of the single-channel images corresponding to other colors is not affected by the single-channel images corresponding to the color shift color, thereby improving the detection and recognition accuracy of fixed pattern noise.

[0063] Furthermore, the fixed pattern noise detection method of the present invention determines whether there is fixed pattern noise in all single-channel images based on the dispersion of column pixel values ​​and row pixel values ​​of all single-channel images, rather than based on traditional edge detection values. Even when the fixed pattern noise does not appear as a whole row or a whole column of stripes, it can still be detected and identified based on the dispersion, which can improve the detection and identification accuracy of fixed pattern noise.

[0064] Example 2

[0065] As an optimization of Embodiment 1, in this embodiment, such as Figure 2 As shown, in step 400, the step of determining whether the test image contains fixed pattern noise based on the single-row pixel values ​​and single-column pixel values ​​of all single-channel images is as follows:

[0066] Step 410: Group all single-row pixel values ​​and single-column pixel values ​​of each single-channel image to obtain multiple row data groups and column data groups. Each row data group has A adjacent single-row pixel values, and each column data group has A adjacent single-column pixel values, where A≥2.

[0067] In step 410, all single-row pixel values ​​of each single-channel image are divided into multiple row data groups, each row data group having A adjacent single-row pixel values, and all single-column pixel values ​​of each single-channel image are divided into multiple column data groups, each column data group having A adjacent single-column pixel values.

[0068] Since the number of pixel rows and columns in a single-channel image is not necessarily the same, the number of row data groups is not necessarily the same as the number of column data groups.

[0069] Preferably, there are no duplicate single-row pixel values ​​among all row data groups, and there are no duplicate single-column pixel values ​​among all column data groups.

[0070] Step 420: Calculate the average of all row data groups and column data groups to obtain multiple row group averages and column group averages.

[0071] In step 420, the pixel values ​​of all single rows in each row data group are summed and then divided by A to obtain the average value of each row data group. One average value of a row data group corresponds to one row data group. The pixel values ​​of all columns in each column data group are summed and then divided by A to obtain the average value of each column data group. One average value of a column data group corresponds to one column data group.

[0072] Step 430: Calculate the difference between the average values ​​of the row groups and the average values ​​of the column groups of three adjacent groups to obtain multiple differences between the row groups and the column groups.

[0073] In step 430, the difference between each row group is 2*R'(middle) + R'(front) + R'(back), where R'(front), R'(middle), and R'(back) are the average values ​​of the first, middle, and last row groups among the three adjacent row group averages, respectively. The difference between the columns is 2*C'(middle) + C'(front) + C'(back), where C'(front), C'(middle), and C'(back) are the average values ​​of the first, middle, and last column groups among the three adjacent column group averages, respectively.

[0074] Preferably, when calculating the row grouping difference and column grouping difference, the average value of each row grouping and the average value of each column grouping are not used repeatedly.

[0075] Step 440: Filter out the maximum value of all row grouping differences and column grouping differences to obtain the maximum difference of row grouping and column grouping.

[0076] In step 440, the differences between all row groups of each single-channel image are compared, and the maximum value is selected as the maximum difference between row groups of that single-channel image. Similarly, the differences between all column groups of each single-channel image are compared, and the maximum value is selected as the maximum difference between column groups of that single-channel image.

[0077] Step 450: Repeat steps 410-440 until the maximum difference between row groups and column groups of all single-channel images is obtained.

[0078] Step 460: Calculate the row discrete value and column discrete value of the test image based on the maximum difference between row groups and the maximum difference between column groups of all single-channel images.

[0079] In step 460, the row discrete value is calculated based on the maximum difference of row groups of all single-channel images, such as the standard deviation or mean difference of the maximum difference of all row groups, and the column discrete value is calculated based on the maximum difference of column groups of all single-channel images, such as the standard deviation or mean difference of the maximum difference of all column groups.

[0080] Step 480: Determine whether the test image contains fixed pattern noise based on the row discrete value, column discrete value and preset discrete threshold.

[0081] In step 480, when determining whether the test image has fixed pattern noise, if the row discrete value or column discrete value is greater than a preset discrete threshold, it is determined that the test image has fixed pattern noise; if both the row discrete value and the column discrete value are less than the preset discrete threshold, it is determined that the test image does not have fixed pattern noise.

[0082] Example 3

[0083] As an optimization of Embodiment 2, in this embodiment, such as Figure 3 As shown, between step 460 and step 480, the following steps are also included:

[0084] Step 470: Repeat steps 410-460 to obtain multiple row discrete values ​​and column discrete values ​​of the test image when the single-channel image is grouped with different A values.

[0085] At this point, in step 480, when determining whether the test image has fixed pattern noise, if one of the discrete values ​​of all rows or one of the discrete values ​​of all columns is greater than a preset discrete threshold, then it is determined that the test image has fixed pattern noise; if all the discrete values ​​of all rows and columns are less than the preset discrete threshold, then it is determined that the test image does not have fixed pattern noise.

[0086] Preferably, the A value used each time the single-channel image is grouped is a positive integer multiple of the A value used in the previous grouping of the single-channel image, and the number of cycles is about 3-5.

[0087] Example 4

[0088] As another optimization of Embodiment 1, in this embodiment, such as Figure 4 As shown, in step 400, the step of determining whether the test image contains fixed pattern noise based on the single-row pixel values ​​and single-column pixel values ​​of all single-channel images is as follows:

[0089] Step 410: Calculate the row discrete value and column discrete value of each single-channel image based on all single-row pixel values ​​and single-column pixel values.

[0090] In step 410, the row discrete value is calculated based on all row pixel values ​​of each single-channel image, such as the standard deviation or average deviation of all row pixel values, and the column discrete value is calculated based on all column pixel values ​​of each single-channel image, such as the standard deviation or average deviation of all column pixel values.

[0091] Step 420: Determine whether there is fixed pattern noise in each single-channel image based on the row discrete value, column discrete value and preset discrete threshold.

[0092] In step 420, when determining whether there is fixed pattern noise in each single-channel image, if the row discrete value or column discrete value is greater than a preset discrete threshold, it is determined that there is fixed pattern noise in each single-channel image; if both the row discrete value and the column discrete value are less than the preset discrete threshold, it is determined that there is no fixed pattern noise in each single-channel image.

[0093] Step 430: Repeat steps 410 and 420 to determine whether fixed pattern noise exists in all single-channel images.

[0094] In step 430, if the image sensor uses a Bayer array, it is necessary to sequentially determine whether there is fixed pattern noise in a red channel image, a blue channel image, and two green channel images.

[0095] Step 440: Based on the noise levels of all single-channel images, determine whether the test image contains fixed pattern noise.

[0096] In step 440, when determining whether the test image has fixed pattern noise, the test image is determined to have fixed pattern noise as long as one of the single-channel images is determined to have fixed pattern noise. Only when all single-channel images are determined to be free of fixed pattern noise is the test image determined to be free of fixed pattern noise.

[0097] Example 5

[0098] A fixed pattern noise detection device for a camera module includes a processor and a memory electrically connected to the processor. The memory stores a computer program for the processor to execute. When the processor executes the computer program, it performs the fixed pattern noise detection method described in Embodiment 1, Embodiment 2, Embodiment 3, or Embodiment 4.

[0099] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the embodiments of the present invention and not to limit them. Although the embodiments of the present invention have been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the embodiments of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting fixed pattern noise in a camera module, characterized in that, Includes the following steps: Step 100: Obtain test images of the camera module; Step 200: Decompose the test image into multiple single-channel images according to the number of channels in the test image; Step 300: Calculate the single-row pixel value of each row of pixels and the single-column pixel value of each column of pixels in each single-channel image. The single-row pixel value is obtained by adding the brightness values ​​of all pixels in the corresponding pixel row and then dividing by the number of pixels in the corresponding pixel row. The single-column pixel value is obtained by adding the brightness values ​​of all pixels in the corresponding pixel column and then dividing by the number of pixels in the corresponding pixel column. Step 400: Determine whether the test image contains fixed pattern noise based on the single row pixel value and single column pixel value of all single channel images; In step 400, the step of determining whether the test image contains fixed pattern noise based on the single-row pixel values ​​and single-column pixel values ​​of all single-channel images is as follows: Step 410: Group all single-row pixel values ​​and single-column pixel values ​​of each single-channel image to obtain multiple row data groups and column data groups. Each row data group has A adjacent single-row pixel values, and each column data group has A adjacent single-column pixel values, where A≥2. Step 420: Calculate the average of all row data groups and column data groups to obtain multiple row group averages and column group averages; Step 430: Calculate the difference between the average row grouping and the average column grouping of three adjacent rows, and obtain multiple row grouping differences and column grouping differences; Step 440: Filter out the maximum value of all row grouping differences and column grouping differences to obtain the maximum difference of row grouping and column grouping; Step 450: Repeat steps 410-440 until the maximum difference between row groups and column groups of all single-channel images is obtained; Step 460: Calculate the row discrete value and column discrete value of the test image based on the maximum difference of all row groups and the maximum difference of all column groups of all single-channel images. The row discrete value is calculated based on the maximum difference of all row groups of all single-channel images, and the column discrete value is calculated based on the maximum difference of all column groups of all single-channel images. Step 480: Determine whether the test image contains fixed pattern noise based on the row discrete value, column discrete value and preset discrete threshold.

2. The fixed pattern noise detection method according to claim 1, characterized in that, In step 480, when determining whether the test image has fixed pattern noise, if the row discrete value or column discrete value is greater than a preset discrete threshold, it is determined that the test image has fixed pattern noise; if both the row discrete value and the column discrete value are less than the preset discrete threshold, it is determined that the test image does not have fixed pattern noise.

3. The fixed pattern noise detection method according to claim 1, characterized in that, Between step 460 and step 480, the following steps are also included: Step 470: Repeat steps 410-460 to obtain multiple row discrete values ​​and column discrete values ​​of the test image when the single-channel image is grouped with different A values.

4. The fixed pattern noise detection method according to claim 3, characterized in that, In step 480, when determining whether the test image has fixed pattern noise, if one of the discrete values ​​of all rows or one of the discrete values ​​of all columns is greater than a preset discrete threshold, then the test image is determined to have fixed pattern noise; if all the discrete values ​​of all rows and columns are less than the preset discrete threshold, then the test image is determined not to have fixed pattern noise.

5. The fixed pattern noise detection method according to claim 1, characterized in that, The test images were captured by the camera module in a dark environment.

6. The fixed pattern noise detection method according to claim 5, characterized in that, During shooting, increase the brightness gain value of the camera module.

7. A fixed pattern noise detection device for a camera module, comprising a processor and a memory electrically connected to the processor, the memory storing a computer program for execution by the processor, characterized in that, When the processor executes the computer program, it performs the fixed pattern noise detection method as described in claim 1.

Images