CN110986775A

CN110986775A - Method for obtaining volume of stratified liquid of centrifugal sample

Info

Publication number: CN110986775A
Application number: CN201911308654.XA
Authority: CN
Inventors: 亓立峰
Original assignee: Shanghai Nao Biotechnology Co Ltd
Current assignee: Shanghai Nao Biotechnology Co Ltd
Priority date: 2019-12-18
Filing date: 2019-12-18
Publication date: 2020-04-10

Abstract

The invention discloses a method for acquiring the volume of stratified liquid of a centrifugal sample, which relates to the technical field of image recognition and comprises the following steps: the mechanical gripper clamps the centrifuged test tube to move to an image acquisition station, the camera acquires a test tube picture, the position of a liquid level interface of the centrifuged test tube is detected and identified by means of an image processing algorithm, and the volumes of different liquids are calculated according to the identified result. The invention realizes automatic sampling of the centrifuged sample and replaces heavy manual test operation.

Description

Method for obtaining volume of stratified liquid of centrifugal sample

[ technical field ] A method for producing a semiconductor device

The invention relates to the technical field of image recognition, in particular to a method for acquiring the volume of stratified liquid of a centrifugal sample.

[ background of the invention ]

Centrifugation is very extensive in medical research and hospital detection aspect application, and the separation of blood mainly separates out the immiscible liquid of different density, and the layering phenomenon is looked over to the manual work to current common, takes a sample to different aspect liquid is perfect, perhaps passes through the laser scanning method, scans from top to bottom, finds out the layering coordinate according to the different principle of the absorption degree of different colour liquid to the laser, then automatic sampling. The following problems exist in manually checking the samples: the work repeatability is high, the attention is reduced due to long-time test sampling, and the misoperation rate is increased along with the increase of the test series. The laser scanning method has the following problems: the measurement speed is slow due to the mechanical movement required. Laser can cause injury to human eyes, and has biological potential safety hazard.

In view of the above, there is a need for an improved sample identification method for centrifugal sample tubes, which overcomes the problems of the prior art.

[ summary of the invention ]

In order to solve the problems, the invention provides a centrifugal sample stratified liquid volume acquisition method without manual identification and laser scanning.

In order to achieve the purpose, the invention adopts the following technical scheme:

a method for obtaining a stratified liquid volume of a centrifuged sample, comprising the steps of:

collecting the image of the centrifuged test tube, preprocessing the image, identifying the preprocessed image after preprocessing, obtaining the pixel position of each liquid layering interface in the test tube, and calculating the volume of each liquid layering according to the pixel position of each liquid layering interface.

Optionally, the preprocessing the image includes removing noise from the acquired image by using a smoothing method.

Optionally, the preprocessed image is identified by using a YOLO algorithm to identify the pixel position of each liquid layer interface, and then the actual position of each liquid layer interface is calculated according to the pixel height of the image.

Optionally, the calculation formula of the actual position of the liquid layer interface is as follows:

wherein, hz_nFor the nth liquid layer interface from top to bottom, hcm is the actual height corresponding to the entire image height, and hs is the entire image pixel height.

wherein s is the cross-sectional area of the test tube.

Optionally, a miniature industrial camera is used to capture images of the centrifuged tube.

Optionally, the centrifuged test tube is grabbed to an image acquisition station by a manipulator, and the image acquisition station is located in the closed space.

Optionally, a white backlight light source is disposed in the enclosed space and used for supplementing light during image acquisition.

The method provided by the invention has the following beneficial effects:

the method provided by the invention realizes automatic sampling of the centrifuged sample, replaces heavy manual test operation, reduces the workload of detection researchers, and simultaneously avoids the situations of misjudgment and misoperation caused by human fatigue and inattention. And moreover, an image recognition mode is adopted, laser scanning is not needed, the recognition speed of the centrifugal sample is improved, the human eyes cannot be injured, each layer of coordinate value of the layered sample is the coordinate of layered liquid relative to the bottom of the test tube, and the influence of different positions of the test tube clamped by the mechanical gripper on the obtained layered liquid coordinate is avoided.

These features and advantages of the present invention will be disclosed in more detail in the following detailed description and the accompanying drawings. The best mode or means of the present invention will be described in detail with reference to the accompanying drawings, but the present invention is not limited thereto. In addition, the features, elements and components appearing in each of the following and in the drawings are plural and different symbols or numerals are labeled for convenience of representation, but all represent components of the same or similar construction or function.

[ description of the drawings ]

The invention will be further described with reference to the accompanying drawings in which:

FIG. 1 is a flow chart of an embodiment of the present invention;

FIG. 2 is a schematic illustration of a centrifuged sample in an embodiment of the invention.

[ detailed description ] embodiments

The technical solutions of the embodiments of the present invention are explained and illustrated below with reference to the drawings of the embodiments of the present invention, but the following embodiments are only preferred embodiments of the present invention, and not all embodiments. Based on the embodiments in the implementation, other embodiments obtained by those skilled in the art without any creative effort belong to the protection scope of the present invention.

Reference in the specification to "one embodiment" or "an example" means that a particular feature, structure or characteristic described in connection with the embodiment itself may be included in at least one embodiment of the patent disclosure. The appearances of the phrase "in one embodiment" in various places in the specification are not necessarily all referring to the same embodiment.

Examples

As shown in fig. 1, the present embodiment provides a centrifugal sample stratified liquid volume acquisition method. Centrifugation is the separation of immiscible liquids of different densities, and the centrifuged sample is shown in FIG. 2, in this example, a 3-layer liquid centrifuged sample is used. The method for obtaining the volume of the stratified liquid of the centrifugal sample comprises the following steps:

the test tube after centrifugation is snatched to the image acquisition station by the manipulator, and the image acquisition station is located centrifugal sample and image acquisition equipment's airtight space, and after the equipment closed shell, image acquisition station internal seal, the collection of image does not receive the environment that external light source influenced.

The miniature industrial camera collects images of the test tube after centrifugation, a white backlight light source is arranged in the closed space, light supplement is carried out in the image collection process, and the condition that the collected images cannot be identified is avoided.

Collecting the image of the centrifuged test tube, setting image recognition regions ROI1 and ROI2 on the image, preprocessing the image, removing noise by adopting a smoothing method, and recognizing the preprocessed image by adopting a YOLO target detection algorithm after preprocessing. The ROI1 and ROI2 are required for calibration work of the YOLO target detection algorithm before training, that is, during calibration, the ROI1 and ROI2 are labeled on a large number of images, specifically, in this embodiment, the positions of the liquid layered interface are framed, and then the model learns the image characteristics of the liquid layered interface by training the algorithm model.

A large number of samples marked with ROI1 and ROI2 position frames and different types of liquid layering interfaces are adopted for training to obtain a trained YOLO target detection model, the process of once inference is the process of performing CNN convolution on an image to extract the image, a group of default positions are generated through anchors, a preselected frame with the largest activation value and probability is selected as final output through weight calculation in a YOLO target detection model network, a category frame of two output liquid layering interfaces is obtained, and finally detailed analysis can be performed according to the category frame of the liquid layering interfaces to obtain the position of an actual liquid level.

The YOLO target detection algorithm and the training algorithm model both belong to the prior art, and are not described herein again.

In this embodiment, a convolution kernel is used to predict the class scores and offsets of a series of default bounding boxes. The principle of the YOLO algorithm is to divide an input image into S × S grid cells, extract image features by multilayer convolution to obtain S × S (B × 5+ C) feature vectors, express the feature vectors in a matrix form, perform regression of a target position frame and an operation of suppressing a non-maximum value, and obtain target frame position and target category information after a series of operations. The object identification and positioning algorithm based on the deep neural network has strong robustness, can still stably decode even if encountering strong light interference, and is quick to operate.

And identifying the pixel position of each liquid layering interface, and calculating the actual position of each liquid layering interface according to the pixel height of the image. The actual position calculation formula of the liquid stratification interface is as follows:

This example uses a centrifuged sample with 3 layers of liquid, and the interface between the liquid and the non-liquid part is calculated, so that from top to bottom the actual position of the top layer of the first layer of liquid, i.e. the interface between the first layer of liquid and the air in the tube, is

The actual position of the interface between the first layer of liquid and the second layer of liquid is

The actual position of the interface between the second layer of liquid and the third layer of liquid is

The actual position of the third layer of liquid bottom layer, namely the interface between the third layer of liquid and the outside of the test tube is

After the pixel position of each liquid layering interface is obtained and the actual position is obtained through calculation, the volume of each liquid layering is calculated, and the actual position calculation formula of each liquid layering interface is as follows:

wherein s is the cross-sectional area of the test tube.

From top to bottom, the height of the first layer of liquid is the top layer of the first layer of liquid, i.e. the difference between the actual position of the interface between the first layer of liquid and the air in the test tube and the actual position of the interface between the first layer of liquid and the second layer of liquid, so the calculation formula for the actual position of the liquid layer interface adopts h_n-h_n+1The method comprises the following steps:

volume of upper layer liquid:

volume of middle layer liquid:

volume of lower layer liquid:

the method provided by the embodiment realizes automatic sampling of the centrifuged sample, replaces heavy manual test operation, reduces the workload of detection researchers, and avoids the situations of misjudgment and misoperation caused by human fatigue and inattention. And moreover, an image recognition mode is adopted, laser scanning is not needed, the recognition speed of the centrifugal sample is improved, the human eyes cannot be injured, each layer of coordinate value of the layered sample is the coordinate of layered liquid relative to the bottom of the test tube, and the influence of different positions of the test tube clamped by the mechanical gripper on the obtained layered liquid coordinate is avoided.

While the present invention has been described with reference to the preferred embodiments, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims. Any modification which does not depart from the functional and structural principles of the present invention is intended to be included within the scope of the claims.

Claims

1. A method for obtaining the stratified liquid volume of a centrifugal sample is characterized in that: the centrifugal sample stratified liquid volume acquisition method comprises the following steps:

2. The short text topic determination method of claim 1, wherein: preprocessing the image includes removing noise from the acquired image using a smoothing method.

3. The intelligent capacitor pre-warning method as claimed in claim 1, wherein: and identifying the preprocessed image by adopting a YOLO algorithm to identify the pixel position of each liquid layering interface, and calculating the actual position of each liquid layering interface according to the pixel height of the image.

4. The method for obtaining a stratified liquid volume of a centrifuged sample as defined in claim 3, wherein: the actual position calculation formula of the liquid stratification interface is as follows:

5. The method for obtaining a stratified liquid volume of a centrifuged sample as defined in claim 4, wherein: the actual position calculation formula of the liquid stratification interface is as follows:

wherein s is the cross-sectional area of the test tube.

6. The method for obtaining a stratified liquid volume of a centrifuged sample as defined in claim 1, wherein: the image of the centrifuged tube was collected using a miniature industrial camera.

7. The method for obtaining a stratified liquid volume of a centrifuged sample as defined in claim 1, wherein: the test tube after centrifugation is grabbed to the image acquisition station by the manipulator, the image acquisition station is located airtight space.

8. The method for obtaining a stratified liquid volume of a centrifuged sample as defined in claim 7, wherein: and a white backlight light source is arranged in the closed space and used for supplementing light in the image acquisition process.