A blood vessel image enhancement and segmentation method and system for a puncture robot

Abstract

The present invention belongs to the medical field and provides a blood vessel image enhancement and segmentation method and system for a puncture robot, which obtains a plurality of continuous original ultrasonic echo signals; maps the plurality of original ultrasonic echo signals into a grayscale image sequence; obtains a first distance sequence between the puncture needle tip and the probe on the grayscale image and a corresponding needle tip grayscale value sequence; fits the first distance sequence and the needle tip grayscale value sequence into a first curve; adjusts the blood vessel detection threshold at different positions on the grayscale image according to the first curve; determines the blood vessels on the grayscale image according to the adjusted blood vessel detection threshold, enhances the brightness of the determined blood vessels, and segments the blood vessels.

Description

Technical Field

[0001] The present invention belongs to the medical field, and in particular relates to a blood vessel image enhancement and segmentation method and system for a puncture robot. Background Art

[0002] Conventional puncture robots often rely on conventional, human-friendly ultrasound images for guidance. However, these ultrasound images undergo complex processing primarily to enhance the physician's visual experience, making them clearer, with higher contrast, and easier to understand and interpret. These images undergo a series of enhancement processes, such as denoising, smoothing, and contrast adjustment, to help physicians quickly identify tissue boundaries, blood vessels, and other anatomical structures.

[0003] While these processing techniques are helpful to doctors, conventional ultrasound images contain many unnecessary variations for puncture robots, which may even affect the accuracy of the robotic system. The puncture robot relies on precise image data to locate the needle tip, navigate the path, and detect target tissue in real time. However, during conventional ultrasound image processing, some details may be overly simplified or smoothed, resulting in information that is irrelevant to the robot's operation being enhanced, while key signals (such as the weak echo characteristics of tiny blood vessels) are suppressed or blurred.

[0004] For example, noise removal and dynamic range compression in ultrasound images often weaken grayscale variations at different depths, making it difficult for robots to accurately perceive the true reflection intensity between layers. While these variations are beneficial to human vision, they can increase misjudgment of target tissue by puncture robots. Furthermore, image smoothing blurs edges, impairing the precise identification of critical structures such as needle tips and blood vessels, thus affecting the accurate positioning and path planning of robotic systems.

[0005] Therefore, while conventional ultrasound images enhance the physician's visual experience, they contain redundant information and processed features that are not needed by the robot. This makes it easy for the robot to be misled or lose sight of key details when performing a puncture based on these images, thus compromising the accuracy and safety of the procedure. Existing technology still has room for improvement in this regard: providing puncture robots with more accurate ultrasound image data that preserves the original signal details, thereby enhancing their autonomous decision-making and operational capabilities. Summary of the Invention

[0006] In order to solve the problems in the prior art, the present invention provides a blood vessel image enhancement and segmentation method for a puncture robot, the method comprising the following steps:

[0007] Acquiring multiple continuous original ultrasonic echo signals;

[0008] mapping the plurality of original ultrasonic echo signals into a grayscale image sequence;

[0009] Acquire a first distance sequence between the puncture needle tip and the probe on the grayscale image and a corresponding grayscale value sequence of the needle tip;

[0010] Fitting the first distance sequence and the needle tip gray value sequence into a first curve;

[0011] adjusting the blood vessel detection thresholds at different positions on the grayscale image according to the first curve;

[0012] The blood vessels on the grayscale image are determined according to the adjusted blood vessel detection threshold, the brightness of the determined blood vessels is enhanced, and the blood vessels are segmented.

[0013] Another aspect of the present invention provides a blood vessel image enhancement and segmentation system for a puncture robot, comprising the following modules:

[0014] An acquisition module, used for acquiring a plurality of continuous original ultrasonic echo signals;

[0015] a mapping module, configured to map the plurality of original ultrasonic echo signals into a grayscale image sequence;

[0016] A calculation module, configured to obtain a first distance sequence between the puncture needle tip and the probe on the grayscale image and a corresponding grayscale value sequence of the needle tip;

[0017] A fitting module, configured to fit the first distance sequence and the needle tip gray value sequence into a first curve;

[0018] an adjustment module, configured to adjust the blood vessel detection thresholds at different positions on the grayscale image according to the first curve;

[0019] The processing module is used to determine the blood vessels on the grayscale image according to the adjusted blood vessel detection threshold, enhance the brightness of the determined blood vessels, and segment the blood vessels.

[0020] The proposed vascular image enhancement and segmentation method for a puncture robot significantly improves the discernibility of blood vessels and surrounding tissue by processing raw ultrasound echo signals and generating grayscale images. Compared to traditional methods, this method directly maps the raw ultrasound signal, avoiding excessive image reconstruction and processing steps. This method preserves subtle differences in the signal and enhances the ability to identify complex tissues. This is particularly critical for the precise positioning of puncture robots during medical procedures.

[0021] By fitting the distance sequence and grayscale value sequence of the needle tip position, a first curve is generated and the vessel detection threshold is dynamically adjusted. This method maintains a fixed ratio of vessel grayscale to needle tip grayscale at different depths, thereby improving the consistency and accuracy of vessel detection. Furthermore, by setting upper and lower thresholds, interference signals from non-vascular areas are eliminated, achieving accurate vessel segmentation. Brightness enhancement and morphological processing are used to further optimize vessel visualization. This method provides more accurate image input for puncture robots, improving the safety and reliability of medical procedures. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. 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 any creative work.

[0023] Figure 1 is a flow chart of the method of the present invention;

[0024] Figure 2 It is the first distance diagram. DETAILED DESCRIPTION

[0025] Below, the invention is preferably described with reference to the accompanying drawings and specific embodiments.

[0026] This embodiment solves the above problem through the following steps:

[0027] In one embodiment, reference Figure 1 The present invention provides a method for enhancing and segmenting blood vessel images using a puncture robot. The puncture robot in this invention refers to a robotic device used in medical procedures that can perform punctures under a doctor's control or automatic control. The robot precisely controls the movement and position of the puncture needle to achieve accurate puncture of the target area, assisting in medical procedures.

[0028] First, a plurality of continuous original ultrasonic echo signals are acquired.

[0029] Ultrasonic waves are continuously emitted through an ultrasonic probe, and echo signals in the target tissue are received through an ultrasonic receiving device. The ultrasonic echo signals are signals reflected from different depths and positions within the target tissue or area, and the signals contain information such as the density and structure inside the tissue. These echo signals are collected in a continuous form by the receiving device, which ensures the temporal and spatial consistency of the ultrasonic signal and can provide accurate tissue information for subsequent image processing steps. In the present invention, the continuous refers to maintaining the coherence of the ultrasonic signal and the continuity of the data at a sufficiently high sampling rate within a certain time period to ensure that the dynamic changes in the target tissue can be fully described. The multiple original ultrasonic echo signals refer to the initial echo data without any processing or transformation, which are used to construct a grayscale image sequence.

[0030] The plurality of original ultrasound echo signals are mapped into a grayscale image sequence.

[0031] Through signal processing and image conversion algorithms, the received continuous ultrasound echo signals are mapped to corresponding grayscale levels according to their intensity, generating a two-dimensional grayscale image sequence. The process includes the following steps:

[0032] First, the raw ultrasound echo signals acquired by the ultrasound probe are preprocessed to extract their reflection intensity information. Next, the amplitude or energy value of each signal is converted to a grayscale value based on a preset mapping function, where higher signal intensities correspond to brighter grayscale levels and lower signal intensities correspond to darker grayscale levels. This mapping rule is used to accurately reflect the density, boundaries, and internal structure of the target tissue in the grayscale image.

[0033] For example, suppose the reflection intensity of an acquired ultrasound echo signal at a specific moment is 250 units. Based on the predefined mapping rules, this reflection intensity is mapped to a grayscale value of 200 (within the grayscale range of 0-255). Consequently, the corresponding pixel in the grayscale image will appear as a brighter grayscale point. An adjacent ultrasound echo signal with an intensity of 50 units, mapped to a grayscale value of 20, will appear as a darker area. Through this continuous signal processing, the grayscale images generated frame by frame will form an image sequence, forming a complete two-dimensional image representation of the target tissue.

[0034] The process of directly mapping multiple raw ultrasound echo signals into grayscale images simplifies the complex operation of processing echo signals into human-friendly visual images in the traditional imaging process. This direct mapping omits the complex reconstruction and rendering process of the signal, converts the signal intensity directly into grayscale values ​​in proportion, and immediately generates an image that can be analyzed and processed by the algorithm. Specifically, in this method, the intensity data of the ultrasound echo signal is not subjected to conventional processing such as interpolation, noise reduction or morphological operations after acquisition. Instead, the signal value is directly mapped to the pixel value in the grayscale image. At the same time, direct mapping retains the subtle differences in the ultrasound echo signal, ensuring that these detailed information, such as slight changes in signal intensity, can be utilized in the subsequent analysis process, enhancing the ability to recognize complex tissue structures.

[0035] A first distance sequence between the puncture needle tip and the probe on the grayscale image and a corresponding grayscale value sequence of the needle tip are obtained.

[0036] Because ultrasound images are fan-shaped and the probe is located at the image's vertex, the pixel distribution in the grayscale image fan-outs from the probe's vertex. The puncture needle typically appears as a linear structure in the grayscale image, with the needle tip at its front end.

[0037] In order to identify the needle tip position, an image processing algorithm suitable for sector images, such as polar coordinate transformation, can be used to convert the curved structure of the sector image into a linear structure suitable for detection.

[0038] The first distance sequence is a radial distance sequence of the pointer tip relative to the ultrasound probe (i.e., the vertex of the image). In a sector image, each frame of the image can be represented as a point in a polar coordinate system, where the probe is located at the origin. Figure 2 As shown, the position of the needle tip can be determined by calculating the radial distance r of its polar coordinates, where r is the distance between the needle tip at the probe vertex and the needle tip. In each frame of the image, the radial distance r is calculated. i (Units are pixels or millimeters), thus forming a radial distance sequence of the needle tip changing with time or image frame number. This radial distance sequence is used to record the depth change of the puncture needle tip throughout the puncture process.

[0039] Assume that in a certain frame of image, the polar coordinates of the needle tip are (r, θ), where r represents the radial distance and θ represents the angle. In the present invention, the focus is on the change of the radial distance and the change of the angle is ignored.

[0040] The needle tip grayscale value sequence refers to the grayscale value sequence corresponding to the needle tip's location in each grayscale image frame. In an ultrasound system, the signal reflection intensity is inversely proportional to the square of the distance. Therefore, during the puncture process, the needle tip's grayscale value in the image will continuously change as the puncture progresses.

[0041] In the sector grayscale image, the needle tip is located at a specific position (r,θ) in polar coordinates. By obtaining the grayscale value g of this point i , i.e., the grayscale intensity value of the pixel where the needle tip is located, forms a sequence of needle tip grayscale values ​​corresponding to each image frame. The needle tip grayscale value reflects the reflection intensity of the ultrasound echo signal at the needle tip. This sequence of needle tip grayscale values ​​in multiple consecutive image frames records the change in ultrasound reflection intensity at the needle tip over time.

[0042] In ultrasound sector images, the radial distance sequence of the needle tip and the grayscale value sequence of the needle tip are correlated, recording the distance of the needle tip relative to the probe and its grayscale changes during puncture. As the needle tip penetrates into different tissue layers, the radial distance r and grayscale value g change synchronously. This correspondence provides accurate data support for subsequent image processing algorithms, particularly for adjusting the threshold for vessel detection.

[0043] Because puncture needles are typically made of metal, they produce significant ultrasonic reflection signals. Therefore, the needle's echo signal is relatively strong in ultrasound images, and the needle body and tip typically appear as brighter areas in grayscale images. This strong signal reflection makes the needle easier to identify in the image, and the converted grayscale image is more reliable. This invention will use the needle tip grayscale as a reference for subsequent operations.

[0044] Furthermore, the present invention uses the Hough transform to identify needle tips. The Hough transform is a common method for detecting straight lines or curves in images and is suitable for identifying objects that appear as straight lines in images. When identifying a puncture needle, the Hough transform can identify the straight line structure of the needle body, while the needle tip can be determined by finding the endpoints of the line.

[0045] The first distance sequence and the needle tip gray value sequence are fitted into a first curve.

[0046] A first distance sequence of the needle tip relative to the probe, identified through ultrasound images, is obtained. This sequence represents the pixel distance of the needle tip in each image frame. Simultaneously, a sequence of needle tip grayscale values ​​corresponding to this first distance sequence is obtained. This grayscale value sequence represents the reflection intensity of the needle tip in each grayscale image frame, reflecting the echo signal strength at the needle tip's location.

[0047] Next, a fitting algorithm is used to fit the two sequences to generate a first curve reflecting the relationship between needle tip position and grayscale value. The fitting algorithm can use linear fitting, polynomial fitting, exponential fitting, or other suitable data fitting methods based on actual needs. The optimal fit ensures that the fitted curve accurately describes the change in the needle tip's distance from the probe and the grayscale value during the puncture process.

[0048] Specifically, each value r in the first distance sequencei The value g in the corresponding needle tip gray value sequence i Using a data fitting algorithm, we can create a curve representing the relationship between needle tip distance and grayscale value. This first curve can intuitively reflect the change in the intensity of the needle tip reflection signal as the needle tip position (distance from the probe) changes during the puncture process.

[0049] The generated first curve can be used to further adjust the threshold setting in image processing, thereby dynamically optimizing the blood vessel detection and enhancement process according to the depth and reflection intensity of the needle tip.

[0050] Example: If the first distance sequence of the needle tip is [r1,r2,...,r n ]The tip gray value sequence is [g1,g2,...,g n ] then (r1,g1),(r2,g2),...,(r n ,g n ) as the fitting point, the first curve g(r) generated by the selected fitting algorithm can be expressed as:

[0051] g(r)=f(r);

[0052] Among them, f(r) is the fitted function, which describes the relationship between the tip distance and the grayscale value.

[0053] The blood vessel detection thresholds at different positions on the grayscale image are adjusted according to the first curve.

[0054] A first curve is obtained by fitting the first distance sequence with the needle tip grayscale value sequence. This curve reflects the grayscale value variation of the needle tip at different depths. To ensure a constant proportional relationship between the grayscale of the blood vessels and the grayscale of the needle tip throughout the image, the variation trend of the needle tip grayscale value with depth in the first curve is analyzed, and a dynamic threshold for blood vessel detection is set based on this trend.

[0055] The specific steps are as follows:

[0056] Calculate the blood vessel detection threshold ratio: set a fixed ratio k between the blood vessel grayscale value and the needle tip grayscale value, that is:

[0057]

[0058] Among them, g vessel is the grayscale value of blood vessels, g needle is the gray value of the needle tip, k is a fixed proportional constant. Dynamically calculate the gray value of the needle tip at different positions: According to the first curve, the gray value g of the needle tip needle (r) is the function of the tip changing with the depth r. For different depths r, the gray value g of the tip at that position is calculated by the first curve.needle (r).

[0059] Determine the blood vessel detection threshold: According to the needle tip gray value g needle (r) and the set proportional constant k, calculate the blood vessel detection threshold T(r) at the corresponding depth position. The blood vessel detection threshold T(r) should satisfy the following relationship:

[0060] T(r)=k·g needle (r);

[0061] This formula ensures that at different depths r, the blood vessel detection threshold is always in a fixed proportion to the grayscale value of the needle tip.

[0062] Dynamically adjust the vessel detection threshold at each location: In each frame of the grayscale image, the grayscale value of the needle tip corresponding to that location is calculated using the first curve based on the distance to the needle tip (i.e., the pixel position in the first distance sequence). Then, the aforementioned ratio formula is applied to dynamically determine the vessel detection threshold at that depth. This ensures that the grayscale value ratio between the blood vessel and the needle tip remains constant at all depths, helping to maintain consistency in image processing.

[0063] Applying the adjusted vessel detection threshold: Use the calculated threshold T(r) to detect and enhance blood vessels at various locations in the grayscale image. Because the vessel detection threshold is based on the needle tip's reflection intensity and a fixed proportional relationship, it ensures consistent contrast between the blood vessel and the needle tip at various depths, improving detection reliability.

[0064] Through the above method, the blood vessel detection threshold is adjusted based on the first curve, ensuring that the ratio of blood vessel grayscale to needle tip grayscale is fixed, so that blood vessels of different depths in the entire image can be consistently detected, avoiding the threshold deviation problem caused by the change of reflection signal with depth.

[0065] For example:

[0066] Assume that at a certain depth r1, the tip gray value is g needle (r1) = 200, proportional constant k = 0.8, then the blood vessel detection threshold T(r1) is:

[0067] T(r1)=0.8×200=160;

[0068] Similarly, at another depth r2, if the tip gray value is g needle (r2) = 150, then the corresponding blood vessel detection threshold T(r2) is:

[0069] T(r2)=0.8×150=120;

[0070] This method ensures that the blood vessel detection threshold and the needle tip grayscale value maintain a fixed proportional relationship at different depths, thereby improving the consistency and accuracy of blood vessel detection.

[0071] The blood vessels on the grayscale image are determined according to the adjusted blood vessel detection threshold, the brightness of the determined blood vessels is enhanced, and the blood vessels are segmented.

[0072] In each frame of grayscale image, the dynamic blood vessel detection threshold adjusted according to the first curve is applied, where the threshold has upper and lower limits T min (r) and T max (r) to ensure that only vascular structures that meet the grayscale range are detected. The upper and lower thresholds represent the lowest and highest grayscale values ​​of the vascular reflection signal, respectively, and are defined as:

[0073] T min (r)≤g(x,y)≤T max (r);

[0074] Among them, g(x,y) is the gray value of the pixel in the image, T min (r) and T max (r) are the dynamic minimum and maximum vessel detection thresholds associated with the distance r, respectively.

[0075] For each pixel g(x,y), only when its gray value is between the threshold range T min (r) and T max (r), it is determined to be a vascular area. The specific judgment conditions are:

[0076] if T min (r)≤g(x,y)≤T max (r), then (x, y)∈vascular region;

[0077] This step can more accurately exclude interference signals and non-vascular areas by defining upper and lower limits.

[0078] After detecting the blood vessel areas, the grayscale values ​​of these areas are enhanced. The purpose of enhancement is to increase the contrast between the blood vessels and the background, making the blood vessels more prominent in the image. The enhancement method can be linear adjustment or nonlinear adjustment. The specific calculation is:

[0079] g enhanced (x,y)=α·g vessel (x,y);

[0080] Among them, α is the enhancement coefficient, g enhanced (x, y) is the enhanced gray value, g vessel (x, y) is the original grayscale value, which ensures that the grayscale value of the enhanced vascular area is significantly improved while maintaining the grayscale balance of the overall image.

[0081] Segment the vascular region based on the upper and lower thresholds and enhanced brightness of the blood vessels. Use morphological processing or connectivity-based segmentation algorithms to extract the complete vascular structure. Specific operations include:

[0082] Morphological processing: Through operations such as dilation and erosion, the edges of blood vessels are optimized and isolated noise is eliminated.

[0083] Region growing: Based on the grayscale value range of the vascular area, it gradually expands from the seed point to ensure that the entire vascular outline is segmented.

[0084] The segmented blood vessel region can be output as a binary image or an enhanced grayscale image for robot image recognition.

[0085] On the other hand, the present invention also provides a blood vessel image enhancement and segmentation system for a puncture robot, comprising the following modules:

[0086] An acquisition module, used for acquiring a plurality of continuous original ultrasonic echo signals;

[0087] a mapping module, configured to map the plurality of original ultrasonic echo signals into a grayscale image sequence;

[0088] A calculation module, configured to obtain a first distance sequence between the puncture needle tip and the probe on the grayscale image and a corresponding grayscale value sequence of the needle tip;

[0089] A fitting module, configured to fit the first distance sequence and the needle tip gray value sequence into a first curve;

[0090] an adjustment module, configured to adjust the blood vessel detection thresholds at different positions on the grayscale image according to the first curve;

[0091] The processing module is used to determine the blood vessels on the grayscale image according to the adjusted blood vessel detection threshold, enhance the brightness of the determined blood vessels, and segment the blood vessels.

[0092] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

[0093] The prior art mentioned in the above background technology section and specific embodiments section of the present invention can be regarded as part of the present invention and used to understand the meaning of some technical features or parameters.

Claims

1. A blood vessel image enhancement and segmentation system for a puncture robot, characterized in that: The system includes the following modules: An acquisition module, used for acquiring a plurality of continuous original ultrasonic echo signals; a mapping module, configured to map the plurality of original ultrasonic echo signals into a grayscale image sequence; A calculation module, configured to obtain a first distance sequence between the puncture needle tip and the probe on the grayscale image and a corresponding grayscale value sequence of the needle tip; A fitting module, configured to fit the first distance sequence and the needle tip gray value sequence into a first curve; an adjustment module, configured to adjust the blood vessel detection thresholds at different positions on the grayscale image according to the first curve; The processing module is used to determine the blood vessels on the grayscale image according to the adjusted blood vessel detection threshold, enhance the brightness of the determined blood vessels, and segment the blood vessels.

2. The blood vessel image enhancement and segmentation system for a puncture robot according to claim 1, characterized in that: Mapping the plurality of original ultrasonic echo signals into a grayscale image sequence comprises: The amplitude or energy value of each signal is converted into a grayscale value according to a preset mapping function, wherein a higher signal intensity corresponds to a brighter grayscale level and a lower signal intensity corresponds to a darker grayscale level.

3. The blood vessel image enhancement and segmentation system for the puncture robot according to claim 1, characterized in that Use Hough transform to identify the needle tip.

4. The blood vessel image enhancement and segmentation system for a puncture robot according to claim 1, characterized in that: Adjusting the blood vessel detection thresholds at different positions on the grayscale image according to the first curve includes: Set a fixed ratio between the grayscale value of the blood vessel and the grayscale value of the needle tip ,Right now: ； in, is the grayscale value of blood vessels, is the tip gray value, is a fixed constant of proportionality; For different depths , calculate the tip gray value at this position through the first curve ; According to the gray value of the needle tip and the set proportional constant , calculate the blood vessel detection threshold corresponding to the depth position: 。 5. The blood vessel image enhancement and segmentation system for a puncture robot according to claim 1, characterized in that: Determining blood vessels on a grayscale image based on the adjusted blood vessel detection threshold includes: In each frame of grayscale image, the dynamic blood vessel detection threshold adjusted according to the first curve is applied, where the blood vessel detection threshold is proportional to the distance Related dynamic minimum and maximum vessel detection thresholds and ; For each pixel (x, y), only its gray value Between threshold range and When the area is between 0 and 1, it is determined to be a vascular area.

Images