Image fusion precision detection method for multispectral fusion system

Abstract

The invention discloses an image fusion precision detection method for a multispectral fusion system, which comprises the following steps of: firstly, combining test modules, fixing a product to be detected on a two-dimensional turntable, and arranging a collimator in a detection area of the product to be detected; then mark points are selected, images detected by a plurality of detectors of the multispectral fusion product are subjected to image fusion processing and then displayed on a computer displayer, and then the center position and six special points around the center position in a small view field of an imaging area of the displayer serve as the mark points. By rotating the two-dimensional turntable, the mark points on the computer display are respectively overlapped with the cross target wire in the collimator, and then the horizontal and pitching angle values of the turntable are recorded one by one. And according to the recorded data and a corresponding formula, the fusion precision of the images in multiple modes after fusion can be calculated. According to the method, technical indexes such as optical axis consistency, view field consistency and relative image inclination in multispectral fusion are combined by utilizing a multipoint mean value algorithm, so that the quality of a fusion image of a multispectral fusion product is quickly detected and calculated.

Description

Technical Field

[0001] The present invention relates to the field of optical detection, and in particular to a method for detecting image fusion accuracy of a multi-spectral fusion system. Background Art

[0002] As a key technology to improve the visual range of human eyes, fusion image processing technology has the conditions to observe visible light band images and infrared band images at the same time, which greatly improves the visible distance of human eyes under various illumination conditions. In products with fusion image processing technology, the smaller the fusion accuracy value, the higher the quality of the fusion image. Whether it is a handheld device or a head-mounted device, as the quality of the fusion image improves, the optical performance and human-machine efficiency of the whole device will be better. Therefore, ensuring the image fusion accuracy of each detection module is the key and difficulty in the adjustment and detection process of multi-spectral fusion products.

[0003] In order to solve the problem of difficult measurement of fusion image quality and reduce the inspection time during adjustment and assembly, a convenient and fast image fusion accuracy detection method is urgently needed. Summary of the invention

[0004] In view of the above problems, the purpose of the present invention is to provide a method for detecting the image fusion accuracy of a multi-spectral fusion system, which can quickly and conveniently detect the fusion image quality (image fusion accuracy) in the fusion image mode of the multi-spectral detection module of the product.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is: a method for detecting image fusion accuracy of a multi-spectral fusion system, comprising the following steps: (1) Combined test module: The test module includes a product under test, connecting cables, a detection device, a data processing device and a collimator; (2) Selecting marking points: Select marking points based on the field of view size of the original mode fused by the product. When the fields of view are consistent, that is, the difference is small and the human eye cannot distinguish them clearly, select 7 marking points based on any mode; when the fields of view are inconsistent, that is, the difference is large and the human eye can distinguish them clearly, select 7 marking points based on the small field of view mode. The original modes include visible light mode, low light mode and infrared mode; (3) Recording marking points: record the seven marking points as O, A, B, C, D, E, and F. Adjust the detection device and align the seven marking points with the target wire in the parallel light tube in turn. The new positions obtained are recorded as: O1, A1, B1, C1, D1, E1, and F1, which are expressed in two-dimensional coordinates as follows: (x 10 ,y 10 );(x 11 ,y 11 );(x 12 ,y 12 );(x13 ,y 13 );(x 14 ,y 14 );(x 15 ,y 15 );(x 16 ,y 16 ); (4) Mode conversion: Switch to the original mode, and then adjust the detection device to align the 7 marking points recorded in step 3 with the target wire in the collimator in turn. The new positions are recorded as: O2, A2, B2, C2, D2, E2, F2, and are expressed in two-dimensional coordinates as: (x 20 ,y 20 );(x 21 ,y 21 );(x 22 ,y 22 );(x 23 ,y 23 );(x 24 ,y 24 );(x 25 ,y 25 );(x 26 ,y 26 ); (5) The horizontal and elevation image fusion accuracy are expressed as R 水平 and R 俯仰 Indicates that: R 水平 =(|x 10 -x 20 |+|x 12 -x 13 |-|x 22 -x 23 |+|x 15 -x 16 | / 2-|x 25 -x 26 | / 2+y 12 -y 15 -y 22 +y 25 ) / 3 R 俯仰 =(|y 10 -y 20 |+|y 11 -y 13 |-|y 21 -y 23 |+|y 14 -y 16 | / 2-|y 24 -y 26 | / 2+x 11 -x 14 -x21 +x 24 ) / 3 Furthermore, the tested products in the test module include infrared and low-light detector fusion, infrared and visible light detector fusion, infrared and infrared detector splicing fusion, low-light and low-light detector splicing fusion; the detection device includes a two-dimensional turntable, which can directly read the horizontal and pitch angle values ​​through the turntable itself.

[0006] The data processing device includes a computer and a display.

[0007] The product under test is directly fixed on the two-dimensional turntable and connected to the computer through a connecting cable. The detection video of the product under test is directly transmitted to the computer end through the connecting cable and is directly displayed on the monitor. After setting up the collimator and the light source, the product is placed together with the two-dimensional turntable in front of the collimator. After the product is powered on, multiple modes such as infrared, visible light, and fusion can detect the target wire. Then the two-dimensional turntable is rotated to record 7 groups of horizontal and pitch angle values ​​after the 7 marking points are aligned with the target wire during the two-mode detection. The recorded data is calculated according to the corresponding formula to obtain the fusion accuracy value after the fusion of the two or more mode detection images.

[0008] The beneficial effect of the present invention is that the present invention adopts a multi-index equalization calculation method, which can comprehensively consider the optical axis consistency, field of view consistency, relative image tilt and other indicators, and divide the three indicators into 1:1:1 equal parts to measure the fusion accuracy index. By reading 7n groups (n modes are fused, 7 groups for each mode) of horizontal and pitch angle value data, and then calculating according to the corresponding formula, the fusion accuracy index of n modes of fusion can be obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 It is a schematic diagram of the positions of the marking points on the display screen of the display of the present invention; Figure 2 It is a schematic diagram of the horizontal and pitch angle values ​​of the two-dimensional turntable when the display mark point coincides with the collimator target wire in the small field of view mode; Figure 3 It is a schematic diagram of the horizontal and pitch angle values ​​of the two-dimensional turntable when the display mark point coincides with the collimator target wire in the large field of view mode; Figure 4 It is a schematic diagram of the test module composition of the present invention; Figure 5 It is a schematic diagram of the evaluation of the optical axis consistency index according to the O1 and O2 angle values ​​in the multi-index equalization calculation method; Figure 6 It is a schematic diagram of the evaluation of the field of view consistency index according to the angle values ​​of A1, B1, C1, D1, E1, F1, A2, B2, C2, D2, E2, and F2 in the multi-index equalization calculation method; Figure 7 It is a schematic diagram of the evaluation of relative image tilt index according to the angle values ​​of A1, B1, D1, E1, A2, B2, D2, and E2 in the multi-index equalization calculation method. DETAILED DESCRIPTION

[0010] The technical solution of the present invention is clearly and completely described below in conjunction with the accompanying drawings and specific embodiments.

[0011] Reference Figure 1-7 As shown, a method for detecting image fusion accuracy of a multi-spectral fusion system comprises the following steps: (1) Combined test module: The test module includes a product under test 1, a connecting cable 2, a detection device, a data processing device 4 and a collimator 5; (2) Selecting marking points: Select marking points based on the field of view size of the original mode fused by the product. When the fields of view are consistent, that is, the difference is small and the human eye cannot distinguish them clearly, select 7 marking points based on any mode; when the fields of view are inconsistent, that is, the difference is large and the human eye can distinguish them clearly, select 7 marking points based on the small field of view mode. The original modes include visible light mode, low light mode and infrared mode; (3) Recording marking points: record the seven marking points as O, A, B, C, D, E, and F. Adjust the detection device and align the seven marking points with the target wire in the parallel light tube in turn. The new positions obtained are recorded as: O1, A1, B1, C1, D1, E1, and F1, which are expressed in two-dimensional coordinates as follows: (x 10 ,y 10 );(x 11 ,y 11 );(x 12 ,y 12 );(x 13 ,y 13 );(x 14 ,y 14 );(x 15 ,y 15 );(x 16 ,y 16 ); (4) Mode conversion: Switch to the original mode, and then adjust the detection device to align the 7 marking points recorded in step 3 with the target wire in the collimator in turn. The new positions are recorded as: O2, A2, B2, C2, D2, E2, F2, and are expressed in two-dimensional coordinates as: (x 20 ,y 20 );(x 21 ,y 21 );(x 22 ,y 22 );(x 23 ,y 23);(x 24 ,y 24 );(x 25 ,y 25 );(x 26 ,y 26 ); (5) The horizontal and elevation image fusion accuracy are expressed as R 水平 and R 俯仰 Indicates that: R 水平 =(|x 10 -x 20 |+|x 12 -x 13 |-|x 22 -x 23 |+|x 15 -x 16 | / 2-|x 25 -x 26 | / 2+y 12 -y 15 -y 22 +y 25 ) / 3 R 俯仰 =(|y 10 -y 20 |+|y 11 -y 13 |-|y 21 -y 23 |+|y 14 -y 16 | / 2-|y 24 -y 26 | / 2+x 11 -x 14 -x 21 +x 24 ) / 3 Reference Figure 4 The tested products in the test module include infrared and low-light detector fusion, infrared and visible light detector fusion, infrared and infrared detector splicing fusion, low-light and low-light detector splicing fusion; the detection device includes a two-dimensional turntable 3, which can directly read the horizontal and pitch angle values ​​through the turntable itself.

[0012] The data processing device includes a computer and a display.

[0013] The product under test is directly fixed on the two-dimensional turntable and connected to the computer through the connecting cable 2. The detection video of the product under test is directly transmitted to the computer end through the connecting cable and displayed directly on the monitor; after setting up the collimator and the light source, the product is placed together with the two-dimensional turntable in front of the collimator. After the product is powered on, various modes such as infrared, visible light, and fusion can detect the target wire. Then the two-dimensional turntable is rotated to record 7 groups of horizontal and pitch angle values ​​after the 7 marking points are aligned with the target wire during the two-mode detection; the recorded data is calculated according to the corresponding formula to obtain the fusion accuracy value after the fusion of the two or more mode detection images.

[0014] The two-dimensional turntable can directly control its horizontal and pitch rotation through the turntable handwheel, and directly read the corresponding horizontal and pitch angle values ​​of the turntable. The product to be tested is directly fixed on the two-dimensional turntable and connected to the computer through a connecting cable. After setting up the collimator (including black body and visible light source), place the product together with the two-dimensional turntable in front of the collimator. After the product is powered on, switch to multiple modes (infrared, visible light, fusion) to detect the target wire. Then rotate the two-dimensional turntable to record the 7 sets of horizontal and pitch angle values ​​after the 7 marking points are aligned with the target wire during two / multiple mode detection. Calculate the recorded data according to the corresponding formula to obtain the fusion accuracy value after the fusion of the two / multiple mode detection images.

[0015] The original modes (visible light mode, low light mode, infrared mode, etc.) of the product fusion may have different field sizes, but this calculation method is applicable to both the fusion accuracy after fusion of two fields of view of equal size and the fusion accuracy when the field sizes are greatly different. When the fields of view are consistent (the difference is small and cannot be clearly distinguished by the human eye), 7 marking points can be selected based on any mode; when the fields of view are inconsistent (the difference is large and can be clearly distinguished by the human eye), 7 marking points are selected based on the small field of view mode.

[0016] In one embodiment, the field of view of the two detection modes is quite different: the infrared mode has a smaller field of view, while the visible light mode has a larger field of view.

[0017] After the product is powered on, turn on the infrared mode. Figure 1 As shown in the figure, select 7 marking points O, A, B, C, D, E, and F, rotate the two-dimensional turntable to align the 7 marking points with the target wire in turn, and record 7 groups of azimuth and pitch angle values ​​including O1 (x10, y10), A1 (x11, y11), B1 (x12, y12), C1 (x13, y13), D1 (x14, y14), E1 (x15, y15), and F1 (x16, y16). Figure 1Middle: Point O is the optical axis position of the small field of view mode detection screen, Point A is the lower half field of view position of the pitch axis of the small field of view mode detection screen, Point B is the right half field of view position of the horizontal axis of the small field of view mode detection screen, Point C is the center position of the fourth quadrant of the small field of view mode detection screen, Point D is the upper full field of view position of the pitch axis of the small field of view mode detection screen, Point E is the left full field of view position of the horizontal axis of the small field of view mode detection screen, and Point F is the lower right corner position of the second quadrant of the small field of view mode detection screen.

[0018] Then switch to visible light mode, rotate the 2D turntable again to align the mark points on the 7 displays with the target wire in turn, and record 7 groups of azimuth and elevation angle values, including O2 (x20, y20), A2 (x21, y21), B2 (x22, y22), C2 (x23, y23), D2 (x24, y24), E2 (x25, y25), and F2 (x26, y26). At this time, the following data is obtained in Table 1: Table 1 According to the data in Table 1, the data in Table 2 can be calculated: Table 2 According to the data in Table 2 and the following formula, the horizontal and elevation image fusion accuracy of the two modes can be calculated respectively: R 水平 =(α 水平 +β 水平 +θ 水平 ) / 3={α 水平 +(β 半水平 +β 全水平 / 2)+θ 水平} / 3 ={|x 10 -x 20 |+|ω1-ω3|+|ω5-ω7| / 2+(θ1-θ3)} / 3 ={|x 10 -x 20 |+(|x 12 -x 13 |-|x 22 -x 23 |)+(|x 15 -x 16 |-|x 25 -x 26 |) / 2+[(y 12 -y 15 )-(y 22 -y 25 )]} / 3 =(|x 10 -x20 |+|x 12 -x 13 |-|x 22 -x 23 |+|x 15 -x 16 | / 2-|x 25 -x 26 | / 2+y 12 -y 15 -y 22 +y 25 ) / 3 R 俯仰 =(α 俯仰 +β 俯仰 +θ 俯仰 ) / 3={α 俯仰 +(β 半俯仰 +β 全俯仰 / 2)+θ 俯仰} / 3 ={|y 10 -y 20 |+|ω2-ω4|+|ω6-ω8| / 2+(θ2-θ4)} / 3 ={|y 10 -y 20 |+(|y 11 -y 13 |-|y 21 -y 23 |)+(|y 14 -y 16 |-|y 24 -y 26 |) / 2+[(x 11 -x 14 )-(x 21 -x 24 )]} / 3 =(|y 10 -y 20 |+|y 11 -y 13 |-|y 21 -y 23 |+|y 14 -y 16 | / 2-|y 24 -y 26 | / 2+x 11 -x 14 -x 21 +x 24 ) / 3 At this time, the image fusion accuracy value R of the two modes of the tested product can be obtained. 水平 , R 俯仰The smaller the value, the better the fusion effect of two / multiple modes of images.

Claims

1. A method for detecting image fusion accuracy of a multi-spectral fusion system, characterized in that: The following steps are involved: (1) Combined test module: The test module includes a product under test, connecting cables, a detection device, a data processing device and a collimator; (2) Selecting marking points: Select marking points based on the field of view size of the original mode fused by the product. When the fields of view are consistent, that is, the difference is small and the human eye cannot distinguish them clearly, select 7 marking points based on any mode; when the fields of view are inconsistent, that is, the difference is large and the human eye can distinguish them clearly, select 7 marking points based on the small field of view mode. The original modes include visible light mode, low light mode and infrared mode; (3) Recording marking points: record the seven marking points as O, A, B, C, D, E, and F. Adjust the detection device and align the seven marking points with the target wire in the parallel light tube in turn. The new positions obtained are recorded as: O1, A1, B1, C1, D1, E1, and F1, which are expressed in two-dimensional coordinates as follows: (x 10 ,y 10 );(x 11 ,y 11 );(x 12 ,y 12 );(x 13 ,y 13 );(x 14 ,y 14 );(x 15 ,y 15 );(x 16 ,y 16 ); (4) Mode conversion: Switch to the original mode, and then adjust the detection device to align the 7 marking points recorded in step 3 with the target wire in the collimator in turn. The new positions are recorded as: O2, A2, B2, C2, D2, E2, F2, and are expressed in two-dimensional coordinates as: (x 20 ,y 20 );(x 21 ,y 21 );(x 22 ,y 22 );(x 23 ,y 23 );(x 24 ,y 24 );(x 25 ,y 25 );(x 26 ,y 26 ); (5) The horizontal and elevation image fusion accuracy are expressed as R 水平 and R 俯仰 Indicates that: R 水平 =(|x 10 -x 20 |+|x 12 -x 13 |-|x 22 -x 23 |+|x 15 -x 16 | / 2-|x 25 -x 26 | / 2+y 12 -y 15 -y 22 +y 25 ) / 3 R 俯仰 =(|and 10 -and 20 |+|and 11 -and 13 |-|and 21 -and 23 |+|and 14 -and 16 | / 2-|and 24 -and 26 | / 2+x 11 -x 14 -x 21 +x 24 ) / 3。 2. The method for detecting image fusion accuracy of a multi-spectral fusion system according to claim 1, characterized in that: The tested products in the test module include infrared and low-light detector fusion, infrared and visible light detector fusion, infrared and infrared detector splicing fusion, low-light and low-light detector splicing fusion; the detection device includes a two-dimensional turntable, and the horizontal and pitch angle values ​​can be directly read through the turntable itself.

3. The method for detecting image fusion accuracy of a multi-spectral fusion system according to claim 1, characterized in that: The data processing device includes a computer and a display.

4. The method for detecting image fusion accuracy of a multi-spectral fusion system according to claim 3, characterized in that: The product under test is directly fixed on the two-dimensional turntable and connected to the computer through a connecting cable. The detection video of the product under test is directly transmitted to the computer end through the connecting cable and is directly displayed on the monitor. After setting up the collimator and the light source, the product is placed together with the two-dimensional turntable in front of the collimator. After the product is powered on, multiple modes such as infrared, visible light, and fusion can detect the target wire. Then the two-dimensional turntable is rotated to record 7 groups of horizontal and pitch angle values ​​after the 7 marking points are aligned with the target wire during the two-mode detection. The recorded data is calculated according to the corresponding formula to obtain the fusion accuracy value after the fusion of the two or more mode detection images.

Images