Endoscope image enhancement method based on LED illumination

Abstract

The invention discloses an endoscope image enhancement method based on LED illumination, and the method comprises the following steps: 1, collecting an endoscope image set which comprises images with different LED intensities and the same observation distance, and images with the same LED intensity and different observation distances; 2, preprocessing an image set, and establishing a relation model of an image brightness coefficient and LED intensity; a picture set with high image quality is selected, and image brightness features are counted by adopting a block weighted gray average and variance method; 3, inputting an endoscope image frame; 4, calculating an image brightness coefficient and an adjustment target value, inputting the target value into the model, calculating LED adjustment intensity, and performing real-time adjustment; 5, carrying out edge enhancement processing on the image; and 6, outputting the enhanced image frame. According to the invention, the LED intensity is accurately controlled in real time, the image illumination condition is improved, and the noise is reduced, so that the image quality is improved, and the disease diagnosis accuracy of clinicians is improved.

Description

Technical Field

[0001] The invention relates to the technical field of medical image processing, and in particular to a method for endoscopic image enhancement based on LED lighting. Background Art

[0002] Endoscopic technology is a medical imaging technology that uses small incisions or natural orifices to observe internal structures and lesions in a non-invasive manner. Endoscopic image processing technology processes the acquired endoscopic images to improve the image quality, thereby showing the internal structures and lesions more clearly.

[0003] In an environment with insufficient light, the image noise level of CMOS imaging will increase, and relying solely on image algorithms to increase brightness will cause image noise to increase further, thereby damaging the overall quality of the image. On the contrary, when the light is too strong, overexposure is prone to occur when examining tissues at close range, making the tissue area image appear completely white, making it difficult to identify its texture details. During endoscopic examinations or surgeries, by accurately adjusting the brightness of the LED in real time, the reflection of tissues 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 of the invention

[0004] The purpose of the present invention is to propose a method for endoscopic image enhancement based on LED lighting. By establishing a corresponding relationship model between image brightness coefficient and illumination, and combining an image enhancement algorithm, the LED intensity is adjusted in real time, and the image can be stably output with sufficient illumination and low noise. High-quality images, thereby improving the clarity and observability of the image. The specific method is as follows: collect multiple groups of endoscopic image samples, and measure the LED light source intensity corresponding to each group of images; pre-process the image samples, extract the brightness characteristics of the image, and calculate the brightness coefficient; according to the brightness coefficient and LED light source intensity data, use a machine learning algorithm to establish a corresponding relationship model between the brightness coefficient and the illumination intensity; during the endoscopic inspection process, collect endoscopic images in real time, calculate the brightness coefficient of the current image for the real-time collected image and calculate the brightness adjustment target value, input the target value into the model, calculate the LED adjustment intensity, and automatically adjust it to keep the endoscopic image quality stable and optimal.

[0005] The purpose of the present invention is achieved through the following technical solutions: A method for endoscopic image enhancement based on LED illumination comprises the following steps: Step 1: Collect multiple sets of endoscopic images, where the image set includes images with different LED intensities and the same observation distance, and images with the same LED intensities and different observation distances; Step 2: Image preprocessing, establish the relationship model between image brightness coefficient and LED intensity; select a set of pictures with high image quality, and use the block weighted grayscale mean and variance method to perform statistical image brightness features; Step 3: Input the endoscope image sequence; Step 4: Calculate the image brightness coefficient and adjust the LED intensity in real time based on the result; Step 5: Perform edge enhancement on the image; Step 6: Output the enhanced image sequence.

[0006] The purpose of the present invention can be further achieved by the following technical measures: In the aforementioned method for endoscopic image enhancement based on LED illumination, step 1 is to collect multiple sets of endoscopic images, wherein the image set includes image sequences with different LED intensities and different observation distances during endoscopic inspection.

[0007] In the aforementioned method for endoscopic image enhancement based on LED illumination, step 2 preprocesses the image set and uses a multi-layer perception model based on a regression task to establish the relationship between the image brightness coefficient and the LED intensity.

[0008] In the aforementioned method for endoscopic image enhancement based on LED lighting, step 3 inputs the collected original image frames into a processing module.

[0009] In the aforementioned method for endoscopic image enhancement based on LED illumination, 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.

[0010] In the aforementioned method for endoscopic image enhancement based on LED lighting, step 5 performs image edge enhancement on the data, and the enhancement method adopts Sobel operator enhancement.

[0011] In the aforementioned method for endoscopic image enhancement based on LED illumination, step 6 outputs the enhanced image frame.

[0012] Compared with the closest prior art, the technical solution provided by the present invention has the following beneficial effects: The present invention improves the stability of image quality: by real-time evaluation of the image brightness coefficient and dynamic adjustment of the LED light source intensity, it can ensure that the endoscopic image maintains sufficient image illumination and minimal noise under different inspection environments, thereby improving image quality.

[0013] The present invention improves diagnostic accuracy: stable image brightness helps doctors observe tissue details more clearly, thereby improving diagnostic accuracy.

[0014] The present invention reduces the complexity of operation: the present invention realizes automatic brightness adjustment, and reduces the operation burden of doctors during the examination process. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a flow chart of the method for endoscopic image enhancement based on LED illumination.

[0016] Figure 2 It is the image brightness coefficient calculation and LED adjustment method. DETAILED DESCRIPTION

[0017] 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. Please refer to Figure 1-2 .

[0018] like Figure 1 As shown, a method for endoscopic image enhancement based on LED illumination comprises the following steps: Step 1: Collect multiple sets of endoscopic images; the image set includes pictures with different LED intensities and the same observation distance, and image sequences with the same LED intensity and different observation distances; Step 2: Image preprocessing; establish a relationship model between image brightness coefficient and LED intensity, train a multi-layer perception model based on regression task; select a set of pictures with high image quality, and use the block weighted grayscale mean and variance method to statistically analyze image brightness features; Step 3: inputting an endoscope picture frame; performing endoscope image input processing; Step 4: Calculate the image brightness coefficient and the adjustment target value, input the target value into the model, calculate the LED adjustment intensity, and make real-time adjustments; Step 5: Perform edge enhancement processing on the image to strengthen the image texture; Step 6: Output the enhanced image frame.

[0019] Furthermore, the relationship model between the image brightness coefficient and the LED intensity in step 2 is: , where L is the LED intensity, M is the image grayscale mean, and P is a multi-layer perception model based on regression tasks; Furthermore, the grayscale mean value of the image is calculated as follows: the image is divided into 3*3 small blocks, and each block is counted separately. The mathematical expression is as follows: ; in, They are Line The grayscale mean and grayscale variance of the column, is the image at coordinates The gray value at . X is the width of the image, and Y is the height of the image. The overall gray mean and variance of the image are calculated using the weighted summation method. The mathematical expression is as follows: ; in, The overall grayscale mean and variance of the image are respectively, Indicates Line The weights of the columns of image patches.

[0020] Select a set of pictures with high image quality, and use the block weighted grayscale mean and variance method to statistically analyze the image brightness characteristics, including the grayscale mean range. And grayscale variance range .

[0021] Furthermore, in step 4, the image frame is subjected to a block weighted grayscale mean method to calculate the image brightness coefficient and the adjustment target value, the target value is input into the model, the LED adjustment intensity is calculated, and real-time adjustment is performed.

[0022] Furthermore, in step 5, the image edge enhancement operator is a Sobel operator.

[0023] like Figure 2 As shown, the specific calculation method of the image brightness coefficient and the LED adjustment method include the following steps: Step 1: Receive an image , the image block grayscale mean and variance method is used to statistically analyze the image brightness characteristics. Among them, The overall grayscale mean and variance of the image respectively; the overall grayscale mean of the current image The ideal image grayscale mean range Compare, calculate the brightness difference and adjust the target value, as shown in the following expression: ; Step 2: When When it is 0, the image brightness is normal and no adjustment is made; when When it is not 0, the image brightness needs to be adjusted. The LED intensity value that needs to be adjusted is calculated based on the relationship between image brightness and LED intensity: ; Step 3: According to The driver PWM circuit adjusts the LED intensity; Step 4: Complete LED light intensity adjustment.

[0024] The above are only preferred embodiments of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should be regarded as the protection scope of the present invention.

Claims

1. A method for endoscopic image enhancement based on LED illumination, characterized in that: The following steps are involved: Step 1: Collect multiple sets of endoscopic images; the image set includes pictures with different LED intensities and the same observation distance, and image sequences with the same LED intensity and different observation distances; Step 2: Image preprocessing; Establish a relationship model between image brightness coefficient and LED intensity, and train a multi-layer perception model based on regression tasks; select a set of high-quality images, and use the block-weighted grayscale mean and variance method to perform statistical image brightness features; Step 3: Input endoscope image frame; Perform endoscopic image input processing; Step 4: Calculate the image brightness coefficient and the adjustment target value, input the target value into the model, calculate the LED adjustment intensity, and make real-time adjustments; Step 5: Perform edge enhancement processing on the image to strengthen the image texture; Step 6: Output the enhanced image frame.

2. The method for endoscopic image enhancement based on LED illumination according to claim 1, characterized in that: The relationship model between the image brightness coefficient and the LED intensity in step 2 is: , where L is the LED intensity, M is the image grayscale mean, and P is a multi-layer perception model based on regression tasks; The grayscale mean value of an image is calculated as follows: the image is divided into 3*3 small blocks, and each block is counted separately. The mathematical expression is as follows: ; in, They are Line The grayscale mean and grayscale variance of the column, is the image at coordinates The gray value at the position; X is the width of the image, and Y is the height of the image; the overall gray mean and variance of the image are statistically calculated using the weighted summation method. The mathematical expression is as follows: ; in, The overall grayscale mean and variance of the image are respectively, Indicates Line The weight of the column image block; Select a set of pictures with high image quality, and use the block weighted grayscale mean and variance method to statistically analyze the image brightness characteristics, including the grayscale mean range. And grayscale variance range .

3. The method for endoscopic image enhancement based on LED illumination according to claim 1, characterized in that: The step 4 uses the block weighted grayscale mean method to calculate the image brightness coefficient and the adjustment target value, inputs the target value into the model, calculates the LED adjustment intensity, and performs real-time adjustment.

4. The method for endoscopic image enhancement based on LED illumination according to claim 1, characterized in that: The specific calculation method of the image brightness coefficient and the LED adjustment method in step 4 include the following steps: Step 1: Receive an image , the image block grayscale mean and variance method is used to statistically analyze the image brightness characteristics; among them, The overall grayscale mean and variance of the image respectively; the overall grayscale mean of the current image The ideal image grayscale mean range Compare and calculate the brightness difference and adjust the target value , as follows: ; Step 2: When When it is 0, the image brightness is normal and no adjustment is made; when When it is not 0, the image brightness needs to be adjusted. The LED intensity value that needs to be adjusted is calculated based on the relationship between image brightness and LED intensity: ; Step 3: According to The driver PWM circuit adjusts the LED intensity; Step 4: Complete LED light intensity adjustment.

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

Images