Infrared blood vessel signal extraction method and system based on local adaptive signal enhancement

Abstract

The invention discloses an infrared angiography and recognition method and system, and relates to the technical field of medical images. According to the blood vessel recognition method, the blood vessel signal features of the human body are analyzed according to the infrared image information of the target human body, and then the blood vessel signals are subjected to filtering, denoising, white balance, ROI extraction, local adaptive enhancement, binaryzation, edge extraction and blood vessel segmentation, so that the defects in the prior art are overcome, and the blood vessel recognition efficiency is improved. According to the method, extraction and segmentation of complex blood vessel signals are realized, especially blood vessel extraction under the complex conditions that the blood vessel signals are weak, the signal-to-noise ratio is low, interference is large, the blood vessel signals are not straight, and the blood vessels are thick and thin, the blood vessel signal extraction efficiency and accuracy are greatly improved, automatic and intelligent image recognition and segmentation of the blood vessels are realized, and the blood vessel extraction efficiency is improved. And the method becomes a functional basis of blood vessel intelligent and automatic application scenes (such as blood vessel puncture).

Description

Technical Field

[0001] The present invention relates to the field of medical imaging technology, and in particular to a method and system for blood vessel infrared imaging and identification extraction. Background Art

[0002] Vascular infrared imaging technology is a non-invasive medical imaging technology used to display and analyze the vascular network in the human body. This technology mainly uses the ability of infrared light to penetrate human tissue and the absorption characteristics of blood to infrared light. The following are the main background technologies of vascular infrared imaging:

[0003] 1. Infrared spectral imaging Principle: Infrared spectral imaging technology is based on the ability of infrared light (wavelength is approximately between 700 nanometers and 1 millimeter) to penetrate human tissue. Different tissues have different absorption and reflectivity of infrared light, which can be used to distinguish blood vessels from surrounding tissues; Applications: In medical imaging, this property is used to enhance visualization of blood vessels, especially in superficial areas of the skin and soft tissue.

[0004] 2. Light Absorption and Reflection Blood absorption: Hemoglobin in the blood has a higher absorption rate for infrared light of a specific wavelength. This absorption difference is used to distinguish blood vessels from other tissues; Infrared Reflectance: Skin and other tissues reflect infrared light differently than blood, which helps improve image contrast.

[0005] 3. Thermal imaging technology Using thermal radiation: Blood vessels produce weak thermal radiation due to blood flow, and infrared thermal imaging technology can detect these thermal differences; Local temperature changes: Local changes in blood flow will cause changes in temperature distribution. This information can be used to reveal the location and status of blood vessels.

[0006] 4. Image Processing Algorithms Enhancement and filtering: Apply various image processing algorithms, such as enhancement, filtering, and edge detection, to improve the clarity and accuracy of vascular images; 3D Reconstruction: Use advanced algorithms to reconstruct the 3D structure of blood vessels from infrared imaging data.

[0007] 5. Camera and detector technology High-sensitivity detectors: Use detectors that are highly sensitive to infrared light, such as cooled infrared detectors, to improve imaging quality; Multi-wavelength imaging: uses infrared light of different wavelengths for imaging to obtain more vascular information.

[0008] 6. Clinical Application Vein localization: used for clinical vein localization, especially in venipuncture and injection; Hemodynamics: The study of blood flow characteristics, such as blood flow velocity and vascular responsiveness.

[0009] Through the combined application of these technologies, vascular infrared imaging provides physicians with a powerful tool to better understand and treat vascular-related diseases.

[0010] The existing traditional image enhancement methods have the following problems in identifying infrared vascular signals: 1. Weak signals cannot be identified 2. The threshold of blood vessel segmentation cannot be dynamic 3. The segmentation effect of blood vessel signals under different lighting conditions is different, and it cannot adapt to lighting conditions 4. It has an amplifying effect on noise and is easy to misidentify noise points 5. The recognition effect of blood vessels of different thicknesses is good or bad, and there is no adaptiveness 6. The recognition effect of blood vessels with different color information is different and cannot be adaptive Summary of the invention

[0011] In order to solve the above problems, the purpose of the present invention is to provide a method and system for identifying and extracting blood vessels by infrared imaging.

[0012] In order to solve the above technical problems, the present invention adopts the following technical solutions: In a first aspect, the present invention provides a blood vessel infrared imaging system, which is a binocular infrared structured light imaging system targeting blood vessel characteristics.

[0013] In a second aspect, the present invention provides a method for extracting and identifying blood vessels in infrared images.

[0014] Step 1: Full image histogram equalization, white balance, dynamically adjust the brightness of the entire image to make the overall brightness of the image moderate.

[0015] Step 2: Median filtering of the full image adaptive filter window is performed to filter out weak noise signals such as skin surface texture. The filter window is adaptive to the size of the blood vessels, which not only retains the blood vessel signals but also filters out weak interference noise.

[0016] Step 3: Local adaptive signal enhancement algorithm. Because the vascular signal is not necessarily the most obvious globally, but it is definitely the most obvious locally, a local sliding window is used. Pixels with signals higher than (average value plus a small threshold) in the window are considered vascular signal points. The sliding window traverses the entire image to extract the vascular signal in each local window.

[0017] Step 4: Extract ROI (region of interest) in the whole image and extract blood vessel signals in the ROI area. ROI extraction is to extract the area where the blood vessels are located by using the connected area within a dynamic threshold range of the depth image as the area where the blood vessels are located.

[0018] Step 5: Selection of vascular signals Morphological selection and true and false judgment algorithms. Because blood vessels have morphological characteristics such as slenderness and directionality, more accurate vascular signals can be screened out through morphological features.

[0019] Compared with the prior art, the present invention has the following beneficial effects: 1. The enhancement effect is obvious, the signal-to-noise ratio is improved, and the weak blood vessel signal can be enhanced to a binary black and white blood vessel signal 2. Adapt to ambient light and automatically adjust white balance 3. Adapt to the size of blood vessel signals, and can extract blood vessel signals within the range of normal blood vessel thickness. 4. Eliminate interference, especially the texture of the skin surface 5. The signals in non-vascular tissue areas are eliminated, and only the real vascular signals in vascular tissue areas are presented. 6. It is adaptive to the strength of blood vessel signals. Even if the blood vessel signal strength is weak, the signal can be enhanced and extracted through local adaptive filtering. 7. The extraction and segmentation of blood vessels are processed with fully automated algorithms throughout the entire process 8. The full-process method has high operating efficiency, with a single-frame processing time of <40ms, and real-time extraction can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0021] Figure 1 The present invention is a schematic flow chart of the steps for implementing the method.

[0022] Figure 2 Schematic diagram of the region growing method for implementing step four of the method of the present invention.

[0023] Figure 3 This is a schematic diagram of an image of ROI extraction in step 4 of the method of the present invention. Implementation

[0024] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0025] See also Figure 1 As shown, the present invention provides an infrared imaging and blood vessel identification and extraction method, comprising:

[0026] Step 1: Full image histogram equalization, white balance, dynamically adjust the brightness of the entire image to make the overall brightness of the image moderate.

[0027] 1. Full image histogram equalization

[0028] Histogram equalization is a method of enhancing the contrast of an image, especially for images where both the background and foreground are too bright or too dark.

[0029] Principle: By adjusting the histogram of the image to achieve a wider and more uniform intensity distribution, this can increase the global contrast of the image, especially when the contrast of a certain area of ​​the image is low.

[0030] Process: Calculate the histogram of the original image, and then design a mapping function to map the original histogram to a more evenly distributed histogram. This mapping function is usually based on the cumulative distribution function (CDF).

[0031] 2. White balance

[0032] White balance is used to adjust the color temperature of an image to make it look more natural.

[0033] How it works: White may appear different colors (such as bluish or yellowish) under different lighting conditions. The purpose of white balance is to adjust the color of the image so that the white areas look truly white, thus making other colors closer to reality.

[0034] Application: In image processing software, you can usually select different white balance presets, or manually adjust the color temperature and tint to achieve the ideal white balance effect.

[0035] 3. Dynamically adjust the brightness of the entire image

[0036] This process involves dynamically adjusting the overall brightness of an image based on its content, making the image appear to be of a more moderate brightness.

[0037] Method: You can use curves adjustment, levels adjustment, or other tools to dynamically adjust the dark and bright areas of the image. The goal is to increase the dynamic range of the image so that details in both the dark and bright areas are preserved.

[0038] Adaptive adjustment: Some advanced methods, such as Local Contrast Enhancement and HDR (High Dynamic Range), can dynamically adjust the brightness and darkness of local areas of the image to make the entire image look more natural.

[0039]

[0040] Step 2: Median filtering of the full image adaptive filter window to filter out weak noise signals such as skin surface texture. The filter window is adaptive to the size of the blood vessels, which not only retains the blood vessel signals but also filters out weak interference noise.

[0041] 1. Basic principles

[0042] Median filtering: Median filtering is a nonlinear digital filtering technique that replaces the value of each pixel in an image with the median of its neighboring pixel values. This method is particularly effective in removing salt and pepper noise from images.

[0043] Adaptive filter window: In traditional median filtering, the filter window size is fixed, but in adaptive filtering, the window size is dynamically adjusted according to the local features of the image. In the vascular area, the window size is adjusted to a size that can capture the blood vessels; in the non-vascular area, the window size is adjusted to a size that is more suitable for denoising.

[0044]

[0045] 2. Technical Implementation Analyze image features: First analyze the image to identify vascular and non-vascular areas. This can be achieved through edge detection algorithms or deep learning-based image segmentation techniques.

[0046] Dynamically adjust the window: In the vascular area, a larger filter window is used to preserve the vascular structure; in other areas, a smaller filter window is used to better eliminate noise.

[0047] Apply median filtering: Apply a median filter to each pixel with a window size dynamically determined based on the area it is in.

[0048] 3. Advantages

[0049] Preserve important information: Improve image quality and diagnostic accuracy by preserving vascular signals while removing unimportant details such as skin surface texture.

[0050] Noise Removal: Effectively remove random noise from images and improve the usability of images.

[0051] Strong adaptability: The adaptive algorithm makes the filtering process more adaptable to the specific characteristics of the image, which is more flexible and effective than traditional fixed window filtering.

[0052] Step 3: Local adaptive signal enhancement algorithm. Because the vascular signal is not necessarily the most obvious globally, but it is definitely the most obvious locally, a local sliding window is used. Pixels with a signal higher than (the average value plus a small threshold) in the window are considered vascular signal points. The sliding window traverses the entire image to extract the vascular signal in each local window. The core of the local adaptive signal enhancement algorithm is to use a sliding window to traverse the entire image and analyze and process the signal within each local window. For vascular imaging, this means emphasizing those pixels in each window that are brighter or darker than the surrounding area, and these pixels are considered part of the vascular signal.

[0053] Technical steps

[0054] 1. Select the window size: Determine the size of the sliding window based on the size of the target vessel and the resolution of the image. The window should not be too large to avoid including too many non-vascular areas, nor too small to avoid missing vascular signals.

[0055] 2. Sliding window processing:

[0056] The window slides across the image from left to right and from top to bottom.

[0057] At each window position, the average of the pixels within the window is calculated.

[0058] 3. Signal detection and enhancement:

[0059] For each pixel within the window, its value is compared to the local mean plus a preset threshold.

[0060] If the pixel value is above this threshold, it is considered to be part of the blood vessel signal and is enhanced.

[0061] 4. Threshold setting:

[0062] Threshold setting is the key to the algorithm, which determines what degree of local contrast is considered as a vascular signal.

[0063] The threshold can be fixed, or dynamically adjusted based on different areas of the image.

[0064] 5. Vascular signal extraction:

[0065] Traverse the entire image, repeat the above steps, extract the blood vessel signal from each local window, and finally obtain an image that highlights the blood vessel signal.

[0066]

[0067] Algorithm Advantages

[0068] 1. Local adaptation: Able to adjust the degree of signal enhancement according to the characteristics of each local area, suitable for complex or non-uniform backgrounds.

[0069] 2. Emphasize important features: effectively emphasize key structures such as blood vessels while suppressing background noise.

[0070] 3. Flexibility and adaptability: Parameters (such as window size and threshold) can be adjusted according to different applications and goals.

[0071] Step 4: Extract ROI (region of interest) in the whole image and extract blood vessel signals in the ROI area. ROI extraction is to extract the area where the blood vessels are located by using the connected area within a dynamic threshold range of the depth image as the area where the blood vessels are located.

[0072] Extracting ROI (Region of Interest) within the entire image, especially vascular signal extraction in medical imaging, is a complex but critical process. This usually involves using 3D depth images to identify vascular tissue and determining the vascular area through dynamic thresholding. The following is a detailed explanation of this process:

[0073] 1. 3D depth image acquisition

[0074] Use depth imaging techniques such as structured light imaging, LiDAR, or stereo cameras to obtain a 3D depth image of the entire scan area.

[0075] These images provide detailed information about the tissue structure below the skin's surface, particularly the location and distribution of blood vessels.

[0076] 2. Dynamic threshold determination

[0077] Based on the depth information, a dynamic threshold is set to distinguish vascular and non-vascular tissues. This threshold is dynamically adjusted according to the statistical distribution of the depth data.

[0078] Dynamic thresholding can be based on the mean and standard deviation of the depth values ​​of the pixels, or use more complex algorithms to take into account characteristics of the local area.

[0079] 3. ROI extraction

[0080] In the 3D depth image, the region where the blood vessels are located is identified by applying a dynamic threshold.

[0081] The identified blood vessel regions are usually presented as connected regions in the depth image, whose depth values ​​are within a specific threshold range.

[0082] 4. Connected Region Analysis

[0083] Image processing techniques, such as the connected component labeling algorithm, are used to identify and segment the vascular regions.

[0084] The shape, size, and distribution of these regions are analyzed to determine the ROI.

[0085]

[0086] 5. Vascular signal extraction

[0087] Within the determined ROI, the vascular signals are further analyzed and extracted.

[0088] This may include the use of edge detection, image enhancement, and other methods to highlight vascular structures.

[0089] 6. Post-processing

[0090] The extracted vascular signals are post-processed, such as smoothing, denoising, or contrast enhancement, to improve image quality and diagnostic value.

[0091] When necessary, three-dimensional reconstruction can be performed to more intuitively display the vascular structure.

[0092] Step 5: Selection of vascular signals Morphological selection and true and false judgment algorithm. Because blood vessels have morphological characteristics such as slenderness and directionality, more accurate vascular signals can be screened out through morphological features.

[0093] The selection of vascular signals, morphological selection and true and false judgment algorithms are a method for identifying and verifying vascular signals based on morphological features. This method is particularly important in image processing because it relies on specific morphological features of blood vessels (such as slenderness and directionality) to improve the accuracy of recognition. The following is a detailed explanation of this method:

[0094] 1. Identification of vascular morphological features

[0095] Feature analysis: First, the typical morphological features of blood vessels are analyzed, such as length, width, directionality, and curvature. Blood vessels usually appear as elongated and continuous structures with specific directions in the image.

[0096] Morphological operations: Apply morphological operations such as erosion, dilation, opening, and closing to enhance or highlight these morphological features. This helps distinguish between vascular and non-vascular structures.

[0097]

[0098] 2. Vascular signal selection

[0099] Morphology-based filtering: Use morphological criteria to filter structures in the image and select those areas that meet the morphological characteristics of blood vessels. For example, you can select structures whose length and width ratios are within a certain range.

[0100] Regional Enhancement: Further enhance the selected area to improve the clarity and visibility of vascular signals.

[0101] 3. True or false judgment algorithm

[0102] Authenticity verification: Algorithms are applied to verify whether the filtered signals truly represent blood vessels. This can be based on other characteristics of the blood vessels, such as blood flow dynamics, color and texture features.

[0103] False positive elimination: exclude structures that meet morphological characteristics but are not actually blood vessels, such as skin texture, hair or image noise.

[0104] 4. Post-processing and analysis

[0105] Morphological refinement: Morphological refinement technology is used to further refine the vascular structure and retain more subtle vascular branches.

[0106] Quantitative analysis: Quantitative analysis of the screened and verified blood vessels was performed, such as measuring their length, width, and branching pattern.

[0107] 5. Integrate other technologies

[0108] Combination with other techniques: To improve accuracy, this method can be combined with other image processing techniques such as edge detection, texture analysis.

[0109] Multimodal data fusion: When necessary, data from other imaging modalities (such as MRI, CT) can be fused to obtain more comprehensive vascular information.

Claims

1. A method for identifying and extracting blood vessels in infrared images, characterized in that: include: Step 1: Full image histogram equalization, white balance, dynamically adjust the brightness of the entire image to make the overall brightness of the image moderate; ； Step 2: Median filtering of the full image adaptive filter window is used to filter out weak noise signals such as skin surface texture. The filter window is adaptive to the size of the blood vessels, which not only retains the blood vessel signals but also filters out weak interference noise. ； Step 3: Local adaptive signal enhancement algorithm. Because the vascular signal is not necessarily the most obvious globally, but it is definitely the most obvious locally, a local sliding window is used. Pixels with signals higher than (average value plus a small threshold) in the window are considered vascular signal points. The sliding window traverses the entire image to extract the vascular signal in each local window. ； Step 4: Extract ROI (region of interest) in the whole image and extract blood vessel signals in the ROI area. ROI extraction is to extract the connected area within a dynamic threshold range of the depth image as the area where the blood vessel is located through the three-dimensional depth image where the blood vessel tissue is located; ； Step 5: Selection of vascular signals Morphological selection and true and false judgment algorithms. Because blood vessels have morphological characteristics such as slenderness and directionality, more accurate vascular signals can be screened out through morphological features; 。 2. A vascular imaging method according to claim 1, characterized in that: The adaptive filter window median filter will adaptively adjust the size of the filter window according to the estimated image pixel size of the blood vessel.

3. A vascular imaging method according to claim 1, characterized in that: The local adaptive enhancement algorithm determines whether the current pixel value is a signal or background based on the comparison between the signal and the surrounding neighborhood background. The size of the neighborhood background is adaptively selected according to the size of the blood vessels. Thus, the neighborhood enhancement signal is extracted. Global enhancement may identify a large amount of noise as a signal, and the local area enhancement signal will effectively extract the real signal.

4. A vascular imaging method according to claim 1, characterized in that: The ROI extraction. After the depth value of the ROI area center is extracted by combining the infrared image and the binocular structured light depth image, the area growth is performed based on the area center. When the depth value is significantly different from the depth value of the area center, the edge point growth is stopped, thereby effectively extracting the ROI area of ​​the human body.

5. A vascular imaging method according to claim 1, characterized in that: Infrared binocular structured light imaging system. This paper adopts an infrared binocular structured light imaging system, which can effectively image and extract blood vessel images and positions.

Images