A method of endoscope image enhancement based on LED illumination

Abstract

The application discloses a kind of based on LED illumination's endoscope image enhancement method, comprising the following steps:1: acquisition endoscope image set, image set includes different LED intensity, the picture under the same observation distance, and the picture under the same LED intensity, different observation distance;2: image set pre-processing, establish the relationship model of image brightness coefficient and LED intensity;Select high-quality image set, and the statistical image brightness feature is carried out using the method of block weighted gray mean and variance;3: input endoscope image frame;4: calculate image brightness coefficient and calculate adjustment target value, input target value into model, calculate LED adjustment intensity, carry out real-time adjustment;5: image is carried out edge enhancement processing;6: output enhanced image frame.The application improves image illumination condition, reduces noise by real-time accurate control LED intensity, to improve image quality, improve the accuracy of disease diagnosis for clinician.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical image processing, in particular to a method for endoscope image enhancement based on LED illumination. BACKGROUND

[0002] Endoscopy is a medical imaging technique that allows the observation of internal structures and lesions through small incisions or natural orifices in a non-invasive manner. Endoscope image processing technology improves the quality of images by processing the acquired endoscope images, thereby more clearly displaying internal structures and lesions.

[0003] In an environment with insufficient light, the image noise level of CMOS imaging will increase, and relying solely on image algorithms to improve brightness will further increase image noise, thereby damaging the overall quality of the image. On the contrary, when the light is too strong, it is easy to cause overexposure when checking the tissue at close range, making the tissue area image appear all white and making it difficult to identify the texture details. During the endoscopy and live surgery process, by accurately regulating the brightness of the LED in real time, the reflection of the tissue caused by strong light can be effectively reduced, and in the case of insufficient light, the limitations of the image signal processor (ISP) algorithm in adjustment can be compensated. SUMMARY

[0004] The purpose of the present application is to provide a method for endoscope image enhancement based on LED illumination, which can output high-quality images with sufficient light and low noise by establishing a model of the corresponding relationship between image brightness coefficients and light, and combining image enhancement algorithms to adjust the intensity of the LED in real time, thereby improving the clarity and observability of the image. The specific method is as follows: collect multiple groups of endoscope image samples, and measure the corresponding LED light source intensity of each group of images; preprocess the image samples, extract the brightness features of the images, and calculate the brightness coefficients; according to the brightness coefficients and LED light source intensity data, use machine learning algorithms to establish a corresponding relationship model between the brightness coefficients and the light intensity; during the endoscopy process, real-time acquisition of endoscope images, calculation of the brightness coefficients of the real-time acquired images and calculation of the brightness adjustment target value, input of the target value into the model, calculation of the LED adjustment intensity, and automatic adjustment to keep the endoscope image quality stable and optimal.

[0005] The purpose of the present application is achieved by the following technical solutions:

[0006] A method for endoscope image enhancement based on LED illumination, comprising the following steps:

[0007] Step 1: Collect multiple groups of endoscope images, which contain pictures under different LED intensities and the same observation distance, and pictures under the same LED intensity and different observation distances;

[0008] Step 2: image preprocessing, establishing the relationship model of image brightness coefficient and LED intensity; selecting a high-quality image set, and using the block weighted gray mean and variance method to statistically analyze the image brightness characteristics;

[0009] Step 3: input the endoscope picture sequence;

[0010] Step 4: image brightness coefficient calculation, and real-time LED intensity adjustment according to the result;

[0011] Step 5: edge enhancement is performed on the image;

[0012] Step 6: output the enhanced image sequence.

[0013] The purpose of the application can be further achieved by the following technical measures:

[0014] The foregoing LED lighting-based endoscope image enhancement method, step 1 collects multiple groups of endoscope images, and the image set contains image sequences of different LED intensities and different observation distances in endoscopy.

[0015] The foregoing LED lighting-based endoscope image enhancement method, step 2 preprocesses the image set, and uses a multilayer perception model based on a regression task to establish the relationship between the image brightness coefficient and the LED intensity.

[0016] The foregoing LED lighting-based endoscope image enhancement method, step 3 inputs the collected original image frames into the processing module.

[0017] The foregoing LED lighting-based endoscope image enhancement method, step 4 calculates the image brightness coefficient of the input data and calculates the brightness adjustment target value, inputs the target value into the model, calculates the LED adjustment intensity, and performs real-time adjustment.

[0018] The foregoing LED lighting-based endoscope image enhancement method, step 5 performs image edge enhancement on the data, and the enhancement method uses a sobel operator for enhancement.

[0019] The foregoing LED lighting-based endoscope image enhancement method, step 6 outputs the enhanced image frames.

[0020] Compared with the closest prior art, the technical scheme provided by the application has the following beneficial effects:

[0021] The application improves the image quality stability: by real-time evaluation of the image brightness coefficient and dynamic adjustment of the LED light source intensity, the endoscope image can maintain sufficient image illumination and minimum noise in different inspection environments, thereby improving the image quality.

[0022] The present application improves the diagnostic accuracy: stable image brightness helps doctors observe tissue details more clearly, thereby improving the accuracy of diagnosis.

[0023] The present application reduces the operation complexity: the present application realizes automatic brightness adjustment, reducing the operation burden of doctors during the examination process. BRIEF DESCRIPTION OF DRAWINGS

[0024] Figure 1 is a method flowchart of endoscope image enhancement based on LED lighting.

[0025] Figure 2 is an image brightness coefficient calculation and LED adjustment method. DETAILED DESCRIPTION

[0026] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application. Please refer to Figures 1-2 .

[0027] As shown in Figure 1 , a method of endoscope image enhancement based on LED lighting comprises the following steps:

[0028] Step 1: Collecting multiple sets of endoscope images; the image set contains pictures under different LED intensities and the same observation distance, and image sequences under the same LED intensity and different observation distances;

[0029] Step 2: Image preprocessing; establishing a relationship model of image brightness coefficient and LED intensity, training a multi-layer perception model based on regression tasks; selecting a high-quality picture set, and using the block weighted gray mean and variance method to statistically analyze the image brightness characteristics;

[0030] Step 3: Inputting the endoscope picture frame; performing endoscope image input processing;

[0031] Step 4: Calculating the image brightness coefficient and calculating the adjustment target value, inputting the target value into the model, calculating the LED adjustment intensity, and performing real-time adjustment;

[0032] Step 5: Performing edge enhancement processing on the image to strengthen the image texture;

[0033] Step 6: Outputting the enhanced image frame.

[0034] Further, the relationship model of image brightness coefficient and LED intensity in step 2 is Where L is the LED intensity, M is the average gray level of the image, and P is a multilayer sensing model based on a regression task;

[0035] Furthermore, the method for calculating the average grayscale value of an image is as follows: divide the image into 3*3 blocks, and perform separate statistics on each block. The mathematical representation is as follows: ;

[0036] in, They are the first Line number The grayscale mean and grayscale variance of the column. The image is in coordinates The grayscale value at point X is the image width, and Y is the image height. The overall grayscale mean and variance of the image are calculated using a weighted average, as shown mathematically below:

[0037] ;

[0038] in, The overall grayscale mean and variance of the image are respectively. Indicates the first Line number Weighted weights of image blocks.

[0039] A set of high-quality images was selected, and the image brightness characteristics, including the gray-level mean range, were statistically analyzed using a block-weighted gray-level mean and variance method. and grayscale variance range .

[0040] Furthermore, in step 4, the image frame is processed using a block-weighted gray-scale mean method to calculate the image brightness coefficient and the adjustment target value. The target value is then input into the model to calculate the LED adjustment intensity and perform real-time adjustment.

[0041] Furthermore, in step 5, the image edge enhancement operator is the Sobel operator.

[0042] like Figure 2 As shown, the specific calculation method for the image brightness coefficient and the LED adjustment method include the following steps:

[0043] Step 1: Receive an image The image brightness characteristics were statistically analyzed using the mean and variance of image block gray levels. The overall grayscale mean and variance of the image are respectively; the overall grayscale mean of the current image is... With the ideal range of image grayscale mean The comparison is performed to calculate the brightness difference and adjust the target value, as shown in the following expression: ;

[0044] Step 2: When When the value is 0, the image brightness is normal and no adjustment is needed; when... When the value is not 0, the image brightness needs to be adjusted. The required LED intensity value is calculated based on the relationship between image brightness and LED intensity. ;

[0045] Step 3: According to Drive the PWM circuit to adjust the LED intensity;

[0046] Step 4: Complete the LED light intensity adjustment.

[0047] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should be considered within the scope of protection of the present invention.

Claims

1. A method for enhancing endoscopic images based on LED illumination, characterized in that, Includes the following steps: Step 1: Acquire multiple sets of endoscopic images; the image set includes images with different LED intensities and at the same observation distance, as well as image sequences with the same LED intensities but at different observation distances; Step 2: Image preprocessing; A model relating image brightness coefficient to LED intensity was established, and a multilayer perception model based on a regression task was trained. A set of high-quality images was selected, and the image brightness features were statistically analyzed using a block-weighted gray-level mean and variance method. Step 3: Input endoscope image frames; Perform endoscopic image input processing; Step 4: Calculate the image brightness coefficient and the target adjustment value. Input the target value into the model to calculate the LED adjustment intensity and perform real-time adjustment. The specific calculation method for the image brightness coefficient and the LED adjustment method include the following steps: Step 4.1: Receive an image The image brightness characteristics were statistically analyzed using the mean and variance of image block gray levels; among them, and These are the overall grayscale mean and variance of the image; the current overall grayscale mean. With the ideal range of image grayscale mean Compare and calculate the brightness difference. and adjusting target values The following expression: ; ; Step 4.2: When When the value is 0, the image brightness is normal and no adjustment is needed; when... When the value is not 0, the image brightness needs to be adjusted. The required LED intensity value is calculated based on the relationship between image brightness and LED intensity. ; Step 4.3: According to Drive the PWM circuit to adjust the LED intensity; Step 4.4: Complete the LED light intensity adjustment; Step 5: Perform edge enhancement processing on the image to strengthen its texture; Step 6: Output the enhanced image frame.

2. The method for enhancing endoscopic images based on LED illumination according to claim 1, characterized in that: The relationship model between the image brightness coefficient and LED intensity in step 2 is as follows: Where L is the LED intensity, M is the average gray level of the image, and P is a multilayer sensing model based on a regression task; The method for calculating the average grayscale value of an image is as follows: Divide the image into 3*3 blocks, and perform separate statistics on each block. The mathematical representation is as follows: ; ; in, and They are the first Line number The grayscale mean and grayscale variance of the column. The image is in coordinates The grayscale value at point X; X is the width of the image, and Y is the height of the image; the overall grayscale mean and variance of the image are calculated using a weighted average method, and the mathematical expression is as follows: ; ; in, and These are the overall grayscale mean and variance of the image. Indicates the first Line number Weighted weights of column image patches; A set of high-quality images was selected, and the image brightness characteristics, including the gray-level mean range, were statistically analyzed using a block-weighted gray-level mean and variance method. and grayscale variance range .

3. The method for enhancing endoscopic images based on LED illumination according to claim 1, characterized in that: The image edge enhancement operator in step 5 is the Sobel operator.

Images