An improved template matching based object tracking algorithm
By automatically detecting the target center point and scale factor to create a fixed-size template and optimizing the template matching algorithm, the problem of high-precision tracking in complex backgrounds is solved, the amount of computation is reduced, and it is suitable for embedded platforms.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- BEIJING AEROSPACE FEITENG EQUIPMENT TECHNOLOGY CO LTD
- Filing Date
- 2024-12-06
- Publication Date
- 2026-06-09
Smart Images
Figure CN119887836B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to an improved template matching-based target tracking algorithm implementation method, particularly to a typical fixed target tracking method for uncooled infrared and visible light images, belonging to image processing technology in the field of computer vision. Background Technology
[0002] In the field of computer vision, target tracking based on uncooled infrared or visible light images is a key technology with broad application prospects. With the rapid development of video surveillance systems, intelligent transportation, and human-computer interaction, the demand for efficient and accurate target tracking algorithms is increasing daily.
[0003] For structures such as bridges, warehouses, and buildings, it is difficult to completely separate the target from the background in complex natural environments. In such cases, region-based template matching algorithms are usually used for target tracking. Template matching algorithms use the grayscale features of the target region as a basis, which includes some background information. They do not have high requirements for image quality, have strong environmental adaptability and anti-interference ability, and are particularly suitable for target tracking technology in complex scenes.
[0004] Common template matching algorithms include the mean absolute difference (MAD), normalized product correlation (NCC), and maximum nearest neighbor distance (MCD). These algorithms are difficult to achieve ideal matching accuracy and correctness. Currently, these traditional matching methods are greatly limited, especially in terms of adaptive scaling and processing time. Meanwhile, some more advanced tracking algorithms based on multi-feature fusion, correlation filtering, and deep learning have relatively high requirements for the processing power of embedded hardware platforms, which also increases the research and development costs. Summary of the Invention
[0005] The purpose of this invention is to overcome the above-mentioned defects and propose an improved target tracking algorithm based on template matching. By automatically detecting the target size at the target center point and creating a fixed-size template, the time consumption of each frame of the image is consistent throughout the tracking process. This invention improves the applicability of traditional template matching algorithms, takes into account both tracking speed and accuracy, and has low and fixed computational load, low requirements for embedded platforms, and has good application prospects.
[0006] The present invention adopts the following technical solution:
[0007] An improved template-matching-based target tracking method includes:
[0008] Let P be the position of the target center point in the t-th frame image. t The scaling factor is denoted as Scale. t t is an integer greater than or equal to 1;
[0009] When t=1, the target size Rect1 in the first frame image is detected based on the target center point position P1 in the first frame image; the target size Rect1 is compared with the template fixed size MaskSize to determine the initial scale factor Scale1, the image is scaled by Scale1 times and a template Mask is created;
[0010] When t > 1, the following steps S1 to S5 are executed cyclically for each frame of the image:
[0011] S1, Scale the image of frame t. t Within a fixed search area (SearchArea), following the point selection method used to create the template, regions of the same size as the template Mask are extracted at intervals. Normalized cross-correlation is then performed to obtain N predicted locations P. t n ;
[0012] S2, using the N predicted positions P t n Using the central point as the starting point, extract regions of the same size as the template Mask point by point in the surrounding area, perform normalized cross-correlation operation, and obtain the first matching position P. t And the first confidence level R t ;
[0013] S3. Determine the scale factor (Scale) based on the scale change of the target in two adjacent frames. t 'Scale the image of frame t' t Within a fixed search area (SearchArea), following the point selection method used to create the template, regions of the same size as the template Mask are extracted at intervals. Normalized cross-correlation is then performed to obtain N predicted locations P. t ' n ;
[0014] S4, using the N predicted positions P t ' n Using the central point as the starting point, extract regions of the same size as the template Mask point by point in the surrounding area, perform normalized cross-correlation operation, and obtain the second matching position P. t 'and second confidence level R t '.
[0015] S5. Based on the first matching confidence level R t Second confidence level R t 'Determine if the scale factor needs to be updated. If so, update the current scale factor and output the matching position as P.' t Otherwise, the scaling factor remains unchanged, and the output matching position is P. t .
[0016] In the improved template-matching-based target tracking method described above, the method for determining the target size Rect1 and the initial scale factor Scale1 in the first frame image based on the target center point position P1 in the first frame image includes:
[0017] Using the target center point P1 in the first frame image as the center, the target area is locally segmented to obtain the target contour points;
[0018] The minimum bounding rectangle of the target is obtained based on the target contour points. The minimum bounding rectangle is then expanded by 3 to 5 pixels to obtain the target bounding box Rect1.
[0019] If Rect1 is smaller than the fixed template size MaskSize, then Scale1 is initialized to 1.0; otherwise, Scale1 is initialized to the ratio of Rect1 / MaskSize.
[0020] In the above-mentioned improved target tracking method based on template matching, the method of locally segmenting the target region to obtain target contour points with the target center point P1 in the first frame image as the center includes: dividing the region around the target center point P1 into multiple directions on an average basis, determining whether the gradient of each pixel in each direction is greater than a preset threshold with P1 as the starting point, and when the gradient of a certain pixel is greater than the preset threshold and the gradient of the previous pixel is less than or equal to the preset threshold, the pixel is taken as the target contour point.
[0021] In the improved template matching-based target tracking method described above, the area around the target center point P1 is divided into 4 to 8 directions on average; the preset threshold value is 20.
[0022] In the improved template-matching-based target tracking method described above, the method for creating the template mask at t=1 includes:
[0023] The template Mask is meshed, with a template step size of 1 in the center region and a template step size of m in the surrounding region, satisfying m>1. With the target center point P1 as the center, the image is scaled according to the scale factor Scale1, and the image data of the corresponding points are read in sequence to complete the template creation.
[0024] In the improved template-matching-based target tracking method described above, based on the scale factor... t The image was scaled using an open bilinear interpolation method.
[0025] In the improved template matching-based target tracking method described above, the first matching position P is obtained. t And the first confidence level R t The methods include:
[0026] Within a fixed search area SearchArea, regions of the same size as the template Mask are obtained at intervals, and normalized correlation operations are performed with the template Mask to obtain the correlation coefficient matrix.
[0027] The pixels corresponding to the N largest correlation coefficients in the correlation coefficient matrix are taken as P. t n ;
[0028] With N predicted positions P t n Using the template Mask as the center point, acquire region blocks of the same size as the template Mask point by point in the surrounding area, perform normalized cross-correlation operation with the template Mask to obtain the correlation coefficient matrix, and take the corresponding point with the largest correlation coefficient matrix as the first matching position P. t and confidence level R t .
[0029] In the improved template matching-based target tracking method described above, the value of N is 5; the surrounding area is 3*3 pixels.
[0030] In the improved template matching-based target tracking method described above, in step S3, the scale factor Scale is predicted based on the scale change of the target in two adjacent frames. t The scale variation of the target in two adjacent frames ranges from 1.02 to 1.07, and the predicted scale factor is Scale. t 'Based on the current scale factor Scale t Obtained by multiplying by the aforementioned scale change.
[0031] In the improved template matching-based target tracking method described above, the second matching position P is obtained. t 'and second confidence level R t The methods include:
[0032] Scale the image t Within a fixed search area (SearchArea), regions of the same size as the template Mask are obtained at intervals, and normalized correlation operations are performed with the template Mask to obtain a correlation coefficient matrix.
[0033] The pixels corresponding to the N largest correlation coefficients in the correlation coefficient matrix are taken as P. t ' n ;
[0034] With N predicted locations P t ' nUsing the template Mask as the center point, acquire region blocks of the same size as the template Mask point by point in the surrounding area, perform normalized cross-correlation operation with the template Mask to obtain the correlation coefficient matrix, and take the corresponding point with the largest correlation coefficient matrix as the second matching position P. t 'and second confidence level R t '.
[0035] In the improved template-matching-based target tracking method described above, in step S5, based on the first matching confidence R... t Second confidence level R t 'Determine whether the scaling factor needs to be updated, if R...' t 0.95 times that of R is still greater than R t And P t and P t If the position is less than 3 pixels, update the current scale factor and output the matching position as P. t Otherwise, the scaling factor remains unchanged, and the output matching position is P. t .
[0036] A computer program product includes a computer program that, when executed by a processor, implements the steps of the above-described method.
[0037] Compared with the prior art, the present invention has at least the following beneficial effects:
[0038] (1) This invention improves the template creation strategy by reading the original image data to create a template, thus solving the problems in traditional template matching algorithms, such as the inability to balance tracking robustness and accuracy, unstable time consumption, and interference from similar surrounding targets, caused by templates that are too large or too small. When searching for each subsequent frame, a scale factor is introduced to match the current scale image and the predicted scale image respectively. Within the search range, a region block of the same size as the template is taken out with each pixel as the center and normalized with the template to obtain the target point position in the current scale and the predicted next scale image. The latest scale factor is obtained by judging the confidence level, thus solving the tracking drift problem caused by the fixed template size in traditional template matching algorithms when the target size exceeds the template at the end of the tracking. This invention improves the applicability of traditional template matching algorithms, balances tracking speed and accuracy, and has a small and fixed computational load, low requirements for embedded platforms, and has good application prospects.
[0039] (2) The present invention adopts an automatic contour detection method for the target area in the first frame image, which can effectively hit the target area, simplify the actual operation process, avoid the subjectivity of manual selection, and make the algorithm more objective; by combining the automatic contour detection method with manual adjustment, the accuracy of locking the target can be further improved.
[0040] (3) In this embodiment of the invention, when creating a template, it is preferred to use a gridded processing method. The central area of the template is the target position, and fine processing is used to ensure the tracking accuracy of the template matching process. The area around the template is the background position, and large step size sampling data is used to obtain more background range, which can ensure the robustness and stability of the template matching process and solve the problem of tracking jump caused by rapid movement or the presence of similar targets in the background.
[0041] (4) The present invention introduces a scale factor in the matching process, which solves the problem of tracking drift caused by the fixed template size in the traditional template matching algorithm when the target size exceeds the template size at the end of the tracking.
[0042] (5) In the entire template matching and tracking process, the data type of this embodiment is integer, the amount of calculation is small and fixed, and the tracking speed and accuracy are balanced. It has low requirements for embedded platforms and can be used for terminal target tracking of image-guided weapons, with good application prospects. Attached Figure Description
[0043] Figure 1 The flowchart below shows the implementation method of the improved template matching algorithm according to an embodiment of the present invention.
[0044] Figure 2 This is a schematic diagram of target bounding box detection in the first frame image of an embodiment of the present invention;
[0045] Figure 3 This is a schematic diagram of template creation in the first frame image of an embodiment of the present invention, wherein (a) is a schematic diagram of the template point selection order, and (b) is a schematic diagram of the template;
[0046] Figure 4 The diagram shows the tracking effect obtained by the infrared target tracking method of the present invention; where (a) is a schematic diagram of the first frame image, (b) is a schematic diagram of the 300th frame image, (c) is a schematic diagram of the 450th frame image, and (d) is a schematic diagram of the 650th frame image. Detailed Implementation
[0047] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments:
[0048] like Figure 1 As shown, the present invention provides an improved target tracking algorithm implementation method based on template matching, which is an infrared target tracking algorithm implementation method based on feature and grayscale fusion, specifically including the following steps:
[0049] 1. Let P be the position of the center point of the target in the t-th frame of the image. t The scaling factor is denoted as Scale. t t is an integer greater than or equal to 1;
[0050] The target region to be tracked is determined based on the target center point pixel P1(x,y) of the first frame image. The target region is locally segmented to find the target edge. Then, the minimum bounding rectangle of the target is calculated by combining morphological methods. Based on the minimum bounding rectangle, the bounding rectangle is expanded by 3 to 5 pixels to determine the size of the target box Rect1. The target size Rect1 is compared with the template fixed size MaskSize to determine the initial scale factor Scale1.
[0051] The specific method is as follows: Taking the target center pixel P1(x,y) as the center, calculate the surrounding grayscale distribution. Divide the 360-degree directional angle into 4 to 8 directions with 45-degree intervals. Starting from P1, determine whether the gradient of each pixel in each direction is greater than a preset threshold. When the gradient of a pixel is greater than the preset threshold and the gradient of the previous pixel is less than or equal to the preset threshold, that pixel is taken as the target contour point. By judging the points whose gradients in each direction are greater than the threshold (the threshold is set to 20 in this embodiment), an initial contour is obtained. The minimum outer rectangle of the target is calculated to obtain the target bounding box size. If the length and width of the target bounding box are both smaller than the fixed size of the template, the initial scale factor Scale1 is initialized to 1.0; otherwise, the initial scale factor Scale1 is initialized to the ratio of Rect1 / MaskSize.
[0052] II. Create a template Mask based on the target position P1 and the initial scale factor Scale1. The method for creating the template is as follows:
[0053] The template mask is meshed, with a template step size of 1 for the central region and a template step size of m (m>1) for the surrounding regions. Using the target center point P1 as the center, the image is scaled according to a scale factor Scale1, and the image data of the corresponding points is read in a specific order to complete template creation. This embodiment of the invention is based on the scale factor Scale1. t Bilinear interpolation is used when scaling the original image.
[0054] Third, for images not in the first frame, a point-by-point search is first performed within a certain search range. Centered on each point, a region of equal size to the template is extracted and subjected to normalized correlation with the template, following the point selection method used when creating the template. This yields a correlation coefficient matrix. The points corresponding to the N (5 in this implementation) largest correlation coefficients in the matrix are used as the initially predicted target point locations P in that frame. t n Then, using N predicted positions P t n Using the central point as the reference, the surrounding area (3x3 in this embodiment) is used to sequentially acquire the aforementioned region blocks. These blocks are then subjected to normalized cross-correlation with the template Mask to obtain a correlation coefficient matrix. The point with the largest correlation coefficient in the matrix is identified as the precise first matching position P.t And the first confidence level R t .
[0055] Fourth, predict the scale factor based on the scale change of the target in two adjacent frames. t In this embodiment, the scale change rate is taken to be in the range of 1.02 to 1.07, and will be adjusted according to the actual application scenario to predict the scale factor. t 'Based on the current scale factor Scale t Obtained by multiplying the target's scale change between two adjacent frames. Image scaling. t Within the search range, the aforementioned region block is also obtained and correlated with the template Mask to obtain the precise second matching position P. t 'and second confidence level R t '.
[0056] V. Based on the first matching confidence level R t Second confidence level R t 'Determine whether the scaling factor needs to be updated, if R...' t 0.95 times that of R is still greater than R t And P t and P t 'Position close to (P in this embodiment)' t and P t If the position is less than 3 pixels, update the current scale factor and output the matching position as P. t Otherwise, the scaling factor remains unchanged, and the output matching position is P. t .
[0057] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method.
[0058] Example 1
[0059] like Figure 1 As shown, the specific steps in this embodiment are as follows:
[0060] Step 1: First, determine the target region to be tracked and the initial scale factor Scale1 based on the target center pixel position P1(x,y) in the first frame image; Using P1(x,y) as the center, perform local segmentation of the target region, find the target edge, extract the target contour points, calculate the minimum bounding rectangle of the target, and then expand the minimum bounding rectangle by 4 pixels to determine the target bounding box Rect1 size, as shown below. Figure 2As shown. If both the length and width of the target bounding box Rect1 are smaller than the fixed size MaskSize of the template, then Scale1 is initialized to 1.0; otherwise, Scale1 is initialized to the ratio of min(Rect1.width,Rect1.height) / MaskSize.
[0061] Step 2: Create a template mask based on the target location P1 and the initial scale factor Scale1. The template mask is created by meshing it, with a template step size of 1 for the central region and 2 for the surrounding regions. After scaling the image using bilinear interpolation according to the scale factor Scale1, image data of corresponding points are read in a specific order, centered on the target location P1(x,y). Figure 3 As shown in (a), each small square in the figure represents a pixel. Red and black represent the read image data. The pixels are arranged in a straight line to complete the template creation. The template effect is as follows: Figure 3 As shown in (b).
[0062] Step 3: For images that are not the first frame, denoted as the t-th frame (t > 1), find the target position P in the previous frame. t Centered on (x, y), a search is performed at intervals within a certain search range. That is, using each pixel as a center point, a region block of the same size as the template Mask and with the same point selection method as the template creation is extracted. The extracted region block and the template Mask are then subjected to normalized correlation operations to obtain a correlation coefficient matrix; the pixels corresponding to the 5 largest correlation coefficients in the correlation coefficient matrix are taken as P. t n Then, using 5 predicted positions P t n Using the center point as the reference point, within the surrounding 3x3 region, obtain the aforementioned region blocks point by point. Perform normalized cross-correlation calculations with the template Mask to obtain the correlation coefficient matrix. The point with the largest correlation coefficient in the matrix is identified as the precise matching position P. t and confidence level R t .
[0063] The basic principle of the normalized correlation tracking algorithm is as follows: Within a certain search area (or the entire image), a region of equal size to the template is extracted centered at each point I(i,j). Normalized correlation is then performed between this region and the template to obtain a correlation coefficient matrix. The point corresponding to the largest correlation coefficient in the matrix is taken as the location of the target point in the image. The formula for calculating the normalized correlation coefficient is shown below:
[0064]
[0065] In the formula:
[0066] r(s,t): Normalized correlation coefficient at search point (s,t);
[0067] M,N: Template size is M×N;
[0068] T(i,j) is the grayscale value at the midpoint (i,j) of the template. This represents the average grayscale value of the template.
[0069] I(i,j) is the gray value at point (i,j) in the region block centered at point (s,t) in the image. This represents the average grayscale value of the area.
[0070] Step 4: Predict the scale factor based on the rate of change of scale. t The general scale change rate is typically between 1.02 and 1.07, adjusted according to the actual application scenario. In this embodiment, a value of 1.04 is used, which is the predicted scale factor. t 'Based on the current scale factor Scale t Obtained by multiplying the scale change rate. The image is scaled according to the predicted scale factor. t New image data to be matched is obtained by scaling using bilinear interpolation.
[0071] Step 5: Based on the scaled new image data to be matched, similar to Step 3, within a fixed search range, following the point selection method used to create the template, take each search point as the center and extract a region block of the same size as the template Mask. Perform normalized cross-correlation to obtain N predicted positions P. t ' n Then, using N predicted positions P t ' n Using the central point as the starting point, extract regions of the same size as the template Mask from the surrounding 3x3 area, perform normalized cross-correlation operations, and obtain the precise matching position P. t 'and confidence level R t '.
[0072] Step 6: Match confidence R based on the two scales t and R t 'Determine if R t 0.95 times that of R is still greater than R t And P t and P t If the position is less than 3 pixels, update the current scale factor and output the matching position as P. t Otherwise, the scaling factor remains unchanged, and the output matching position is P. t .
[0073] The improved template matching tracking algorithm was completed based on the above steps, and simulation experiments were conducted to verify the results. Figure 4 As shown, (a) is a schematic diagram of the first frame image, (b) is a schematic diagram of the 300th frame image, (c) is a schematic diagram of the 450th frame image, and (d) is a schematic diagram of the 650th frame image. The white box represents the template matching result. It can be seen that the target obtained by the present invention has achieved stable and accurate tracking effect at different scales.
[0074] The above description is only the best specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the protection scope of the present invention.
[0075] The contents not described in detail in this specification are common knowledge to those skilled in the art.
Claims
1. An improved target tracking method based on template matching, characterized in that, include: The first t The position of the target center point in the frame image is denoted as The scaling factor is denoted as , t It is an integer greater than or equal to 1; t When =1, it is based on the target center point position in the first frame image. Detecting the size of the target in the first frame image ; to the target size Fixed size with template Compare and determine the initial scale factor. Scaling the image Double and create template ; t When the value is greater than 1, the following steps S1 to S5 are executed cyclically for each frame of the image: S1, the first t Frame Image Scaling Times, within a fixed search range Within, following the point selection method used to create the template, points are retrieved at intervals that correspond to the template. Perform normalized cross-correlation on regions of the same size to obtain N predicted locations. ; S2, using the N predicted positions Using the template as the center point, extract points one by one from the surrounding area. For regions of the same size, perform a normalized cross-correlation operation to obtain the first matching position. and first confidence level ; S3. Determine the scale factor based on the scale change of the target in two adjacent frames. , will the t Frame Image Scaling Times, within a fixed search range Within, following the point selection method used to create the template, points are retrieved at intervals that correspond to the template. Perform normalized cross-correlation on regions of the same size to obtain N predicted locations. ; S4, using the N predicted positions Using the template as the center point, extract points one by one from the surrounding area. For regions of the same size, perform a normalized cross-correlation operation to obtain the second matching position. Second confidence level ; S5. Based on the first matching confidence level Second confidence level Determine if the scale factor needs to be updated. If so, update the current scale factor and output the matching position. ; Otherwise, the scaling factor remains unchanged, and the output matching position is... ; In step S5, based on the first matching confidence level... Second confidence level Determine whether the scale factor needs to be updated. 0.95 times is still greater than ,and and If the position is less than 3 pixels, update the current scale factor and output the matching position. ; Otherwise, the scaling factor remains unchanged, and the output matching position is... .
2. The improved template-matching-based target tracking method according to claim 1, characterized in that, Based on the target center point position in the first frame image Determine the target size in the first frame image and initial scaling factor The methods include: The target center point position in the first frame image Centered on the target area, the target region is locally segmented to obtain the target contour points; The minimum bounding rectangle of the target is obtained based on the target contour points. This minimum bounding rectangle is then expanded outwards by 3-5 pixels to obtain the target bounding box. ; like Smaller than the template fixed size but Initialize to 1.0, otherwise Initialize to / The ratio of .
3. The improved template-matching-based target tracking method according to claim 2, characterized in that, The target center point position in the first frame image Methods for obtaining target contour points by locally segmenting the target region around a central point include: Position of the target center point The surrounding area is divided into multiple directions on average, with Starting from the point, determine whether the gradient of each pixel in each direction is greater than a preset threshold. When the gradient of a certain pixel is greater than the preset threshold and the gradient of the previous pixel is less than or equal to the preset threshold, the pixel is taken as the target contour point.
4. The improved template-matching-based target tracking method according to claim 3, characterized in that, Position of the target center point The surrounding area is divided into 4 to 8 directions on average; the preset threshold value is 20.
5. The improved template-matching-based target tracking method according to claim 1, characterized in that, t When =1, create template. The methods include: template The grid is processed with a template step size of 1 for the central region and a template step size of m for the surrounding regions, satisfying m>1, with the target center point position as the reference. Centered on the image, scale it according to the scale factor. After scaling, the image data of the corresponding points are read in sequence to complete the template creation.
6. The improved template-matching-based target tracking method according to claim 1, characterized in that, According to the scaling factor The image was scaled using an open bilinear interpolation method.
7. The improved template-matching-based target tracking method according to claim 1, characterized in that, Get the first matching position and first confidence level The methods include: Within a fixed search range Inside, get the interval and template Regions of the same size, and templates Perform normalized correlation operations to obtain the correlation coefficient matrix; The pixels corresponding to the N largest correlation coefficients in the correlation coefficient matrix are taken as... ; With N predicted positions Using the center point as the reference point, obtain the template point by point in the surrounding area. Regions of the same size, and templates Perform normalized cross-correlation to obtain the correlation coefficient matrix, and take the corresponding point with the largest correlation coefficient in the matrix as the first matching position. and confidence level .
8. The improved template-matching-based target tracking method according to claim 7, characterized in that, The value of N is 5; the value of the surrounding area is 3. 3 pixels.
9. The improved template-matching-based target tracking method according to claim 1, characterized in that, In step S3, the scale factor is predicted based on the scale change of the target in two adjacent frames. The scale variation of the target in two adjacent frames ranges from 1.02 to 1.07, and the predicted scale factor is... Based on the current scale factor Obtained by multiplying by the aforementioned scale change.
10. The improved template-matching-based target tracking method according to claim 1, characterized in that, Get the second matching position Second confidence level The methods include: Scale the image Times, within a fixed search range Inside, get the interval and template Regions of the same size, and templates Perform normalized correlation operations to obtain the correlation coefficient matrix; The pixels corresponding to the N largest correlation coefficients in the correlation coefficient matrix are taken as... ; With N predicted positions Using the center point as the reference point, obtain the template point by point in the surrounding area. Regions of the same size, and templates Perform normalized cross-correlation to obtain the correlation coefficient matrix. The point with the largest correlation coefficient in the matrix is then used as the second matching position. Second confidence level .
11. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method of claim 1.