Long-wave infrared image strip noise suppression method and system based on one-dimensional side window filtering
Through one-dimensional side window filtering technology, the long-wave infrared images are decomposed and noise extraction are solved, which solves the problem of strip noise caused by ambient temperature changes, and efficient denoising and miniaturization of infrared cameras is achieved, avoiding the baffle correction method. High cost and power consumption and risk of missed information.
Patent Information
- Application Number
- CN202410965449.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-18
- Publication Date
- 2024-11-22
AI Technical Summary
When the ambient temperature of existing infrared cameras changes, the strip noise will change accordingly, resulting in a decrease in the frame rate of acquisition. The baffle correction method is costly, high power consumption and easy to miss important information, especially in non-destructive testing and autonomous driving tasks. There is a risk of missed inspection.
The method based on one-dimensional edge window filtering is used to decompose high and low frequency information of long-wave infrared images, and through one-dimensional horizontal edge window filtering and one-dimensional column-directional guidance filtering, strip noise is accurately extracted and removed to ensure the retention of vertical edges.
Effectively remove strip noise, reduce cost and power consumption, realize the miniaturization of infrared cameras, improve application capabilities in high frame rate tasks, and avoid information missed detection.
Smart Images

Figure CN119006323A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a method and system for suppressing image stripe noise, and more specifically, to a method and system for suppressing long-wave infrared image stripe noise based on one-dimensional edge window filtering, belonging to the field of image denoising. Background Art
[0002] Existing infrared cameras often use baffles (approximately constant temperature flat black bodies) to correct strip noise. In actual use, the baffle is operated by a servo or other power mechanism, and the average of several frames of baffle infrared images is collected as a correction item. However, strip noise will change (drift) with the change of ambient temperature. When the ambient temperature changes drastically, obvious strip noise will appear in a short time. In this case, the baffle correction method requires frequent interruption of the acquisition process, which will affect the camera's acquisition frame rate, resulting in limitations on its application in tasks that require high frame rates. Secondly, in tasks such as non-destructive testing and autonomous driving (pedestrian detection), the use of baffle correction methods can easily lead to missing information at important time points, causing missed detections. In addition, the power mechanism required to operate the baffle will also bring additional costs and power consumption, which is not conducive to the miniaturization of infrared cameras.
[0003] At present, when using a long-wave infrared camera to collect infrared images, if baffle correction cannot be used, column-wise stripe noise will appear, which appears as column-wise stripes. Therefore, it is an inevitable trend to study a long-wave infrared image stripe noise suppression method and system that can suppress noise while accurately retaining the vertical edges. Summary of the invention
[0004] In order to solve the above-mentioned problems in the prior art, the present invention provides a method and system for suppressing stripe noise in long-wave infrared images based on one-dimensional edge window filtering, which has the technical characteristics of effectively removing stripe noise in long-wave infrared images and accurately retaining vertical edges.
[0005] In order to achieve the above object, the present invention is implemented by the following technical solutions:
[0006] The present invention provides a method for suppressing stripe noise of long-wave infrared images based on one-dimensional edge window filtering, the method comprising the following steps:
[0007] Step 1) inputting an (infrared) image, using a one-dimensional horizontal edge window filter (method) to decompose high- and low-frequency information (information separation) of the (infrared) image containing strip noise, so as to obtain a smooth part and a high-frequency part under the premise of accurately retaining the vertical edge in the smooth part;
[0008] Step 2) accurately extracting the strip noise in the high-frequency part, using the smooth part as a guide image and the high-frequency part as a filter image, and using a one-dimensional column-oriented guide filter (method) to obtain the strip noise;
[0009] Step 3) Subtract the stripe noise from the input (infrared) image to remove the stripe noise, and finally, output the (denoised) image.
[0010] Preferably, the one-dimensional horizontal side window filtering expression in step 1) is:
[0011] q=f(p)
[0012] hg=pq
[0013] Where p is the input infrared image, f(·) is a one-dimensional horizontal edge window filter, · in (·) indicates omission, q is the smoothing part, and h f This is the high frequency portion containing stripe noise and texture.
[0014] Preferably, the one-dimensional column-oriented guided filtering expression in step 2) is:
[0015] s=f CGF (filtering=hf, guidance=q)
[0016] Among them, f CGF (·) is a one-dimensional column-oriented guided filter, where · in (·) means omission, h f is the filtered image, q is the guide image, and S is the strip noise.
[0017] Preferably, in step 3), the expression for subtracting stripe noise from the input image is:
[0018] o=ps
[0019] Among them, o is the output image, S is the strip noise, and p is the input infrared image.
[0020] Preferably, the one-dimensional horizontal side window filtering (method) used in step 1) specifically includes:
[0021] The pixel to be processed is placed on the edge of the filter window (since a one-dimensional horizontal filter window is used, only the left window and the right window need to be considered), and the filter window is 1 / 4 of the image width H (expressed as ), according to the filtering results of the left window and the right window, the filtering result closest to the input is selected as the output of the one-dimensional horizontal edge window filtering (method), so as to maximize the retention of the vertical edge of the smooth part and avoid the confusion of the vertical edge with the strip noise.
[0022] Preferably, the size of the one-dimensional horizontal filter window is set to 1×(2r+1), the left window is defined as the 1st to r+1th pixel points, and the right window is defined as the r+2th to the last pixel point (i.e., the 2r+1th).
[0023] Preferably, in step 2), the strip noise (the output image of the one-dimensional column-wise guided filtering) and the guided image within the window range are in a linear relationship, and the strip noise and the smooth part within a column satisfy a piecewise linear relationship.
[0024] The invention discloses a system for suppressing strip noise of long-wave infrared images based on one-dimensional edge window filtering, comprising a computer with computer program code and a storage medium, wherein the storage medium is one or more of a disk storage, a CD-ROM and an optical storage.
[0025] Beneficial effects: The present invention has the technical characteristics of being able to effectively remove strip noise from long-wave infrared images and accurately retain the edges in the vertical direction; when removing strip noise, no power mechanism is required, which reduces costs and power consumption compared to the prior art; the technical solution of this application provides a basis and direction for realizing the miniaturization development of infrared cameras. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] Figure 1 It is the overall flow chart of the present invention.
[0027] Figure 2 It is a simplified flow chart of the present invention.
[0028] Figure 3 1 is a one-dimensional schematic diagram of a step edge in an embodiment of the present invention. DETAILED DESCRIPTION
[0029] The present invention will be further described below in conjunction with the accompanying drawings, but the present invention is not limited to the following embodiments.
[0030] like Figure 1-3 The figure shows a specific embodiment of a method for suppressing stripe noise in a long-wave infrared image based on one-dimensional edge window filtering. The embodiment is a method for suppressing stripe noise in a long-wave infrared image based on one-dimensional edge window filtering. The method comprises the following steps:
[0031] Step 1) inputting an infrared image, and using a one-dimensional horizontal edge window filtering method to decompose the high-frequency and low-frequency information of the infrared image containing strip noise (i.e., information separation), so as to obtain the smooth part and the high-frequency part under the premise of accurately retaining the vertical edge of the smooth part;
[0032] The one-dimensional horizontal side window filtering expression is:
[0033] q=f(p)
[0034] hf=pq
[0035] Where p is the input infrared image, f(·) is a one-dimensional horizontal edge window filter, · in (·) represents omission, q is the smooth part, and hf is the high-frequency part containing strip noise and texture.
[0036] Step 2) accurately extracting the strip noise in the high-frequency part, using the smooth part as a guide image and the high-frequency part as a filter image, and using a one-dimensional column-oriented guide filtering method to obtain the strip noise;
[0037] The one-dimensional column-oriented guided filtering expression is:
[0038] s=f CGF (filtering=hf, guidance=q)
[0039] Among them, f CGF (·) is a one-dimensional column-oriented guided filter, where · in (·) means omission, h f is the filtered image, q is the guide image, and s is the strip noise.
[0040] Step 3) subtracting the stripe noise from the input infrared image to remove the stripe noise, and finally outputting a denoised image;
[0041] The expression for subtracting stripe noise from the input image is:
[0042] o=ps
[0043] Among them, o is the output image, s is the strip noise, and p is the input infrared image.
[0044] In a preferred embodiment, the one-dimensional horizontal side window filtering method used in step 1) specifically includes:
[0045] The pixel to be processed is placed on the edge of the filter window. Since a one-dimensional horizontal filter window is used, only the left window and the right window need to be considered. The filter window is 1 / 4 of the image width H, which can be expressed as According to the filtering results of the left window and the right window, the filtering result closest to the input is selected as the output of the one-dimensional horizontal edge window filtering method, so as to maximize the retention of the vertical edge of the smooth part and avoid the confusion of the vertical edge with the strip noise.
[0046] In a preferred embodiment, the size of the one-dimensional horizontal filter window is 1×(2r+1), the left window is defined as the 1st to r+1th pixel points, and the right window is defined as the r+2th to the last pixel point (i.e., the 2r+1th pixel points). It should be pointed out that a variety of filtering methods can be used here. In order to simplify the calculation, the simplest mean filter can be used.
[0047] like Figure 3As shown, taking the typical step edge as an example, in the one-dimensional case x 1∶3 and x 4∶6 The pixel points on both sides of the edge have pixel values of l and h respectively. The key point that one-dimensional horizontal edge window filtering can maintain the edge is that, taking point x3 as an example (r=2), the calculation result of the left window is:
[0048]
[0049] The calculation result of the right window is:
[0050]
[0051] After obtaining the calculation results of the left and right windows, select the optimal output, the standard is the closest to the input:
[0052]
[0053] Therefore, the result is Similarly, we can calculate x4 points. (in, ). This shows that there is no blurred edge.
[0054] In a preferred embodiment, in step 2), the strip noise (i.e., the output image of the one-dimensional column-wise guided filtering) within the window range is in a linear relationship with the guided image, while the strip noise within a column satisfies a piecewise linear relationship with the smooth part.
[0055] The present invention discloses a system for a method for suppressing stripe noise in a long-wave infrared image based on one-dimensional side window filtering, comprising a computer with a computer-usable program code and a storage medium, wherein the storage medium is one or more of a disk storage, a CD-ROM, and an optical storage. It should be understood by those skilled in the art that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing a computer-usable program code.
[0056] The present application is described with reference to the flowcharts and / or block diagrams of the methods, systems, and computer program products of the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing device to work in a specific way, so that the instructions stored in the computer-readable memory produce a product including an instruction device, which implements the functions specified in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide for implementing the process in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0057] Finally, it should be noted that the present invention is not limited to the above embodiments, and there are many variations. All variations that can be directly derived or associated with the content disclosed by ordinary technicians in this field should be considered as the protection scope of the present invention.
Claims
1. A method for suppressing stripe noise in long-wave infrared images based on one-dimensional edge window filtering. Features The method comprises the following steps: Step 1) Input an image, and use a one-dimensional horizontal edge window filter to decompose the high- and low-frequency information of the image containing strip noise, so as to obtain the smooth part and the high-frequency part under the premise of accurately retaining the vertical edge of the smooth part; Step 2) accurately extracting the strip noise in the high-frequency part, using the smooth part as a guide image and the high-frequency part as a filter image, and using one-dimensional column-oriented guide filtering to obtain the strip noise; Step 3) Subtract the stripe noise from the input image to remove the stripe noise, and finally output the image.
2. According to claim 1, a method for suppressing stripe noise in long-wave infrared images based on one-dimensional edge window filtering, Features: The one-dimensional horizontal side window filtering expression in step 1) is: q=f(p) hf=pq Where p is the input infrared image, f(·) is the one-dimensional horizontal edge window filter, q is the smooth part, and hf is the high-frequency part containing strip noise and texture.
3. According to the method for suppressing stripe noise in long-wave infrared images based on one-dimensional edge window filtering according to claim 1, Features: The one-dimensional column-oriented guided filtering expression in step 2) is: s=f CGF (filtering=hf,guidance=q) Among them, f CGF (·) is a one-dimensional column-oriented guided filtering, hf is the filtered image, q is the guided image, and s is the stripe noise.
4. According to claim 1, a method for suppressing stripe noise in long-wave infrared images based on one-dimensional edge window filtering, Features: The expression for subtracting stripe noise from the input image in step 3) is: o=ps Among them, o is the output image, s is the strip noise, and p is the input infrared image.
5. The method for suppressing stripe noise in long-wave infrared images based on one-dimensional edge window filtering according to claim 1 or 2, Features: The one-dimensional horizontal side window filtering used in step 1) specifically includes: The pixel point to be processed is placed on the edge of the filtering window, and the filtering window is 1 / 4 of the image width H. According to the filtering results of the left window and the right window, the filtering result closest to the input is selected as the output of the one-dimensional horizontal edge window filtering, so as to maximize the retention of the vertical edge of the smooth part and avoid confusion between the vertical edge and the strip noise.
6. According to claim 5, a method for suppressing stripe noise in long-wave infrared images based on one-dimensional edge window filtering, Features: Assume that the size of the one-dimensional horizontal filter window is 1×(2r+1), the left window is defined as the 1st to r+1th pixel points, and the right window is defined as the r+2th to the last pixel points.
7. The method for suppressing stripe noise in long-wave infrared images based on one-dimensional edge window filtering according to claim 1 or 3, Features: In step 2), the strip noise and the guide image are in a linear relationship within the window range, while the strip noise and the smooth part within a column satisfy a piecewise linear relationship.
8. A system for implementing the method for suppressing stripe noise in long-wave infrared images based on one-dimensional edge window filtering as described in any one of claims 1 to 7, Features: It includes a computer with a computer-usable program code and a storage medium, wherein the storage medium is one or more of a disk storage, a CD-ROM, and an optical storage.
