CN114486687B - Multi-scale continuous observation feedback method and device for femtosecond laser processing cells - Google Patents

Abstract

The multi-scale continuous observation and feedback method and device for femtosecond laser processing of cells of the present invention belong to the field of cell laser processing and detection. The present invention is an ultrafast continuous imaging system based on ultrafast pulse sequences. The time resolution can reach femtosecond level. It can continuously observe biologically specific cell regulation and processing processes in real time, collect cell size and shape at the same time, and use cell image processing to extract cells. The size value is coupled with the time domain broadening method through the collaborative controller to obtain the detection image information within the time scale of nanoseconds to milliseconds during the cell processing process, thereby obtaining a multi-scale observation system and establishing a feedback mechanism for cell regulation and processing-detection. Achieve multi-scale continuous observation, and can achieve rapid detection and precise control of the processing process. The present invention can also use specific images of individual cells to guide the technical optimization of femtosecond laser processing of cells, and realize the application of ultrafast laser-controlled processing of cells with high precision, high efficiency, low damage and controllable biological impact.

Description

Multi-scale continuous observation and feedback method and device for femtosecond laser processing of cells

Technical field

The invention relates to a multi-scale continuous observation and feedback method and device for femtosecond laser processing of cells, particularly a real-time feedback method and device that can realize multi-scale continuous observation of the cell processing process, and belongs to the field of cell processing observation.

Background technique

Using ultrafast laser to process cells can achieve high-precision cell stimulation, perforation and other non-destructive and minimally invasive cell processing processes, and can be widely used in biomedical fields such as stem cell cultivation and gene therapy research. The process of cell processing by femtosecond laser is a non-equilibrium and nonlinear dynamic process. The processing process and mechanism still need to be further studied and discussed. At different time scales, the electronic dynamic evolution process of the interaction between femtosecond laser and cells is different. In addition, the dynamic regulation process of cell processing by parameters such as laser flux and frequency affects the damage, subsequent survival and growth of cells. or the influence of variation plays a key role. Observing the interaction between laser and cells in these different time scales will help explain the processing mechanism and improve processing efficiency. Moreover, the specificity of biological cells determines the non-repeatability of the processing process. Therefore, continuous observation within an ultra-fast time is required to accurately capture key information during the processing process. However, the current laser processing of cells can usually only be observed after processing using methods such as laser confocal imaging or ultrasonic detection. Real-time continuous observation of the femtosecond laser processing of cells is not possible, which seriously restricts the precise control of the cell processing process. .

Contents of the invention

The purpose of the present invention is to solve the problem of difficulty in precise control and rapid detection during femtosecond laser processing of cells, and to provide a multi-scale continuous observation and feedback method and device for femtosecond laser processing of cells, based on ultrafast pulse sequences. The continuous imaging system has a time resolution that can reach femtosecond level. It can continuously observe biologically specific cell regulation and processing processes in real time, collect cell size and morphology at the same time, and use cell image processing to extract cell size values. Through the collaborative controller and time domain The broadening method is coupled to obtain detection image information within the time scale of nanoseconds to milliseconds during cell processing, and a multi-scale observation system is obtained. A feedback mechanism for cell regulation, processing and detection is established to achieve rapid detection and precise control of the processing process. , use specific images of individual cells to guide the technical optimization of femtosecond laser processing of cells, and achieve the application of ultrafast laser-controlled processing of cells with high precision, high efficiency, low damage, and good controllability.

The present invention is achieved through the following technical solutions:

The multi-scale continuous observation and feedback method of femtosecond laser processed cells disclosed by the invention includes the following steps:

Step 1. The ultrafast laser generated by the ultrashort pulse laser passes through the multi-frequency pulse sequence generator and becomes a pulse sequence with femtosecond time delay. It is divided into two beams through beam splitting, one of which passes through the objective lens. It focuses and acts on the cell sample whose cell flow rate is controlled by the microfluidic device to perform non-destructive or minimally invasive regulatory processing of the cells with multi-frequency pulsed femtosecond laser; after another laser pulse sequence passes through the sample, it carries ultrafast information. Received by CCD, ultra-fast continuous images are generated and stored in the computer;

Step 2. Extract the cell images constrained in the microfluidic chip captured by the CCD camera in step 1. Each image contains one or more cells. The cell equivalent diameter value d is calculated through the cell image processing method and output to Synergy. in the controller. The time resolution t of the time domain broadened detection image is calculated to be the reciprocal of the laser repetition frequency f, that is, f=1/t. According to the time span T of the required detection image and the matching relationship between the equivalent cell diameter d and the cell flow rate v: Calculate the cell flow rate v, set the parameters of the microfluidic device, and realize feedback adjustment of the detection process;

Step 3: The cooperative controller triggers the time domain broadening pulse generator to generate a pulse laser with a repetition frequency of f calculated in step 2. After time domain broadening, a laser pulse with any time delay in the range of nanoseconds to milliseconds is generated. After acting on the sample, the image detector and signal generator collect the time-domain broadened detection images of the cells at this time resolution and store them in the computer. Combined with the ultra-fast continuous images received by the CCD camera, multi-scale continuous observation of cells is obtained. Information, observe the biological reaction process of cells, and guide the optimization of femtosecond laser cell processing. The cell biological reaction process includes cell morphology, cell flow, cell membrane perforation, and healing.

The specific implementation method of the cell image processing method in step two is:

(1) Convert the ultra-fast continuous image containing cell morphology in step 1 into a grayscale image, and draw the grayscale histogram of the image;

(2) Select a threshold from the trough of the histogram to perform binary segmentation, and process the image into a binary image;

(3) Fill the holes in the binary image and delete the connecting lines between cells to achieve the purpose of segmenting cell adhesions;

(4) Perform morphological opening operations to improve the visual smoothness of the image;

(5) Calculate the size of the coverage area of each cell and exclude block contours with too small or too large areas;

(6) Then find the center of gravity of the connected domain and draw numbers at the coordinate points of the center of gravity, and finally generate a cell counting marker image;

(7) Output the equivalent cell diameter d.

The invention discloses a multi-scale continuous observation and feedback device for femtosecond laser processing of cells, which is used to implement the multi-scale continuous observation and feedback method of femtosecond laser processing of cells. The multi-scale continuous observation and feedback Devices, including ultrashort pulse lasers, multi-frequency pulse sequence generators, processing objectives, microfluidic devices for controlling the flow rate of cell samples, collaborative controllers, time domain broadening pulse generators, spatial beam splitters, and spatial light combiners , image detector, several beam splitters, CCD camera and computer. The ultrafast pulse laser generated by the ultrashort pulse laser is converted into a pulse sequence consisting of several multi-frequency pulses with time delays through a multi-frequency pulse sequence generator, and is divided into two beams by a beam splitter, one of which passes through the processing objective lens It is used for laser processing of cells. After another beam passes through the sample, the cell signals within the femtosecond time scale of the cell processing process are collected. The pulse lasers of different frequencies carrying imaging information are separated through several beam splitters and entered into the CCD camera for collection. signal, generating ultra-fast continuous images and storing them in the computer. The above image is identified through the cell image processing method to identify the equivalent diameter of the cells and sent to the collaborative controller. Subsequently, the laser pulse repetition frequency of the time domain broadening pulse generator is calculated according to the repetition frequency calculation method in step 2, and adjusted with the required cell flow rate. The parameters of the microfluidic device are used to adjust the cell flow rate. The collaborative controller is used to trigger the time domain broadening pulse generator to generate pulse lasers of different frequencies with time delays in the order of nanoseconds to milliseconds. The spatial spectrometer is used to separate the different frequencies. The pulses are separated and acted on the sample, and signals at any time scale from nanoseconds to milliseconds during cell processing are collected. The spatial light combiner is used to gather the imaging information carried by the pulse laser at different times at the same spatial position, and is detected sequentially through the image. The detector receives the signal, forms a time domain broadened detection image and stores it in the computer. After computer processing, multi-scale cell processing imaging information is obtained, the light stimulation and ablation processes during cell processing are observed, and the cell processing process and results are fed back. The multiple scales include femtoseconds, nanoseconds, microseconds or milliseconds.

The ultrashort pulse laser is an ultrafast laser capable of generating laser light with a single wavelength and a pulse width of femtoseconds.

The multi-frequency pulse sequence generator includes a dispersion device, a delay device, a beam splitter, a reflector, and an objective lens; the dispersion device is a nonlinear crystal or an optical fiber; and the delay device is a quartz sheet or a grating.

The collaborative controller uses the repetition frequency calculation method in step 2 to set the laser repetition frequency of the time domain broadening pulse generator by calculating the relationship between the cell equivalent diameter, the laser repetition frequency of the time domain broadening pulse generator, and the cell flow rate, and controls The frequency of the time-domain broadening pulse acting on cells enables the time resolution of the time-domain broadening detection image to reach nanosecond or millisecond levels.

The time domain broadened pulse generator is a combination of a high repetition frequency pulse laser capable of generating a predetermined center wavelength and a prism to generate broadened laser pulses.

Depending on the time resolution, the image detector is a high-speed camera or a combination of a photoelectric detector and an oscilloscope.

Beneficial effects:

1. The multi-scale continuous observation and feedback method of femtosecond laser processed cells disclosed in the present invention couples the femtosecond laser ultrafast continuous imaging and time domain broadening image detection methods to the femtosecond laser cell processing process, and cooperates with the cell image and laser The coordinated control of frequency can obtain information on the femtosecond laser cell processing process in the nanosecond, microsecond or even millisecond time scale range, and realize real-time multi-scale continuous observation of the femtosecond laser cell processing process.

2. The multi-scale continuous observation and feedback method of femtosecond laser processing cells disclosed in the present invention compensates for the femtosecond time scale problem of non-repeatable cell processing processes by observing the ultrafast dynamic interaction between femtosecond laser and cells at different times. Continuous camera blank.

3. The multi-scale continuous observation and feedback method of femtosecond laser processed cells disclosed in the present invention can obtain the intensity map and high-resolution state image of each cell in real time by accurately controlling the resolution of imaging information, and combine the obtained cell data information The image processing method establishes a feedback mechanism for cell regulation, processing and detection to provide real-time feedback on the processing effect, enabling in-depth research on laser-regulated cell processing methods, timely adjustment of processing parameters, and ultimately achieving ultra-fast processing with high precision, high efficiency, low damage, and controllable biological impact. Application of laser controlled processing of cells.

Description of the drawings

Figure 1 is a schematic flow chart of the multi-scale continuous observation and feedback method of femtosecond laser processing cells of the present invention.

Among them: 1-ultrashort pulse laser, 2-multi-frequency pulse sequence generator, 3-beam splitter, 4-objective lens, 5-cell sample constrained by microfluidic device, 6-9-reflector, 10-CCD Camera, 11—computer, 12—coordinated controller, 13—time domain broadening pulse generator, 14—spatial beam splitter, 15—spatial light beam combiner, 16—image detector.

Detailed ways

In order to better illustrate the purpose and advantages of the present invention, the content of the invention will be further described below in conjunction with the accompanying drawings and examples.

Example 1

In order to realize the observation of cell states and ultra-fast dynamic characteristics at different times during femtosecond laser processing of cells, a multi-scale continuous observation and feedback method of femtosecond laser processing of cells disclosed in this embodiment is used, which includes the following steps:

Step 1. The ultrafast laser generated by the ultrashort pulse laser passes through the multi-frequency pulse sequence generator and becomes a pulse sequence with femtosecond time delay. It is divided into two beams through beam splitting, one of which passes through the objective lens. It focuses and acts on the cell sample whose cell flow rate is controlled by the microfluidic device to perform non-destructive or minimally invasive control and processing of the cells with multi-frequency pulsed femtosecond laser; after another laser pulse sequence passes through the sample, it will carry ultrafast information The light is received by the CCD, which generates ultra-fast continuous images and stores them in the computer;

Step 2: Extract the images of cells constrained in the microfluidic chip captured by the CCD camera in step 1. Each image contains one or more cells. Use the cell image processing algorithm to first convert the image containing cells into a grayscale image, draw the grayscale histogram of the image, and then select a threshold from the trough of the histogram for binary segmentation to process the image into a binary image. Call the hole filling function to fill the holes in the binary image and delete the connecting lines between cells, thereby effectively segmenting cell adhesions. Then perform morphological opening operations to improve the visual smoothness of the image and calculate the coverage area of each cell. size, exclude block outlines that are too small or too large, then find the center of gravity of the connected domain and draw numbers at the center of gravity coordinate points, and finally generate a cell count marker image, complete the entire image processing process and output the equivalent diameter of the cell. The cell equivalent diameter value d calculated by the above cell image processing method is output to the collaborative controller; the time resolution t of the time domain broadening detection image is calculated to be the reciprocal of the laser repetition frequency f. According to the time span T of the required detection image and the matching relationship between the equivalent cell diameter d and the cell flow rate v Calculate the cell flow rate v, set the parameters of the microfluidic device, and realize feedback adjustment of the detection process;

Step 3: The cooperative controller triggers the time domain broadening pulse generator to generate a pulse laser with a repetition frequency of f calculated in step 2. After time domain broadening, a laser pulse with a nanosecond time delay is generated, which is separated by space and time. On the flowing sample, each laser pulse carries a part of the sample imaging information. After spatial beam combination, the image detector and signal generator collect the time domain broadened detection image of the cell at the time resolution and store it in the computer. middle. Combined with the ultra-fast continuous images received by the CCD camera, multi-scale continuous detection information of cells can be obtained, and biological reaction processes such as cell morphology, cell flow, cell membrane perforation, and healing can be directly observed, and the cells exposed to femtosecond laser radiation can be intuitively revealed through images. The morphological changes produced after illumination can analyze the ultra-short continuous evolution process of the interaction between femtosecond laser and cell membranes and organelles.

This embodiment discloses a multi-scale continuous observation and feedback device for femtosecond laser processing cells, which is used to implement the multi-scale continuous observation and feedback method for femtosecond laser processing cells disclosed in this embodiment. The multi-scale continuous observation and feedback device The observation feedback device and its working method are as follows: the ultrafast pulse laser generated by the ultrashort pulse laser 1 is converted into a pulse sequence composed of several multi-frequency pulses with a certain time delay through the multi-frequency pulse sequence generator 2, and is split into beams. Mirror 3 is divided into two beams, one of which passes through the processing objective lens 4 for laser processing of cells, and the other beam passes through the sample 5 to collect cell signals within the femtosecond time scale of the cell processing process, carrying imaging information. Pulse lasers of different frequencies are separated by a number of beam splitters 6-9, enter the CCD camera 10 to collect signals, and store ultra-fast continuous images in the computer 11. Use cell image processing methods to convert the above-mentioned ultra-fast continuous images containing cell morphology into grayscale images, draw the grayscale histogram of the image, and then process the image into a binary image to further segment cell adhesions and improve image vision. Smooth effect, then calculate the size of the coverage area of each cell, generate a cell counting mark image, and finally output the equivalent diameter of the cell. According to the calculation method in step 2, the laser pulse frequency and cell flow rate for controlling the time domain broadening pulse generator 13 are calculated, and the collaborative controller 12 triggers the time domain broadening pulse generator 13 to generate a pulse with a nanosecond-level time delay. Pulse lasers of different frequencies are separated by the spatial beam splitter 14 and applied to the sample 5 to collect signals within the nanosecond time scale of the cell processing process, and the spatial light combiner 15 is used to gather the pulse lasers at different times. The carried imaging information is sequentially received by the image detector 16 at the same spatial position to form a time domain broadened detection image and stored in the computer 11 . After computer processing, multi-scale cell processing imaging information from femtosecond to nanosecond is obtained, and the light stimulation and ablation process during cell processing can be observed. It can also be used to observe biological reactions such as cell morphology, cell flow, cell membrane perforation, and healing. process to guide the optimization of cell processing technology.

The specific implementation process of this embodiment is:

The ultrashort pulse laser with a wavelength of 800nm and a pulse width of 35fs is generated by the pulse laser 1. The ultrashort pulse passes through the multi-pulse sequence generator 2, specifically through the photonic crystal fiber, and the ultrashort pulse with a central wavelength of 800nm is broadened to cover 400nm. The supercontinuum laser pulse in the -1100nm band passes through the dispersion delay dielectric quartz sheet to produce a pulse sequence consisting of four laser pulses with wavelengths of 400nm, 600nm, 800nm, and 1000nm. The time delay between adjacent pulses is 200fs. The above pulse sequence is divided into two paths after passing through the first beam splitter 3. One path is focused on the cell sample 5 through the objective lens 4. The sample 5 is constrained and controlled by the microfluidic device; the other path of the laser pulse sequence acts on the cell sample in a delayed sequence. On sample 5, the cell imaging information carrying different processing times passes through four beam splitters 6-9 and then the four pulse laser beams are spatially separated and imaged through the CCD camera 10 to obtain the cells at laser processing 0fs, 200fs, 400fs and 600fs. The ultra-fast continuous images are stored in the computer 11.

Use the cell image processing method to process the above imaging. Use the cell image processing algorithm to first convert the ultra-fast continuous images containing cells into grayscale pictures, draw the grayscale histogram of the image, and then select a threshold from the trough of the histogram. Binary segmentation, processes the picture into a binary image, calls the hole filling function to fill the holes in the binary image, and deletes the connecting lines between cells, thereby effectively segmenting cell adhesions, and then performs morphological opening operations , improve the visual smoothness of the image, calculate the size of the coverage area of each cell, exclude block contours with too small or too large areas, then find the center of gravity of the connected domain and draw numbers at the center of gravity coordinates, and output the equivalent diameter of a single cell is about 10 μm . In order to obtain cell processing information with a time resolution of 1 ns and a detection image time span of 10 ns, it is calculated that the repetition frequency of the time domain broadening pulse generator should be 10 ⁹ Hz and the cell flow rate of the microfluidic device should be 10 ³ m /s, adjust the cell flow rate of the microfluidic device to the above value. The time domain broadening pulse generator 13 is a combination of a high repetition frequency pulse laser and a prism. It is controlled by the cooperative controller 12 and generates a laser with a repetition frequency of 10 ⁹ Hz and a center wavelength of 800 nm. After dispersion by the prism, the wavelength range is 790nm- The 810nm pulse laser is separated in time, and the broadened pulse laser is separated in spatial position through the spatial beam splitter 14. The spatial beam splitter is selected as a grating and acts on the flowing sample 5. The 810nm laser is the first. After arriving at the sample, the 790nm laser finally reaches the sample. The pulsed laser carries the sample imaging information and enters the spatial light combiner 15 through the necessary reflectors, which converges the pulsed lasers at different spatial locations at different times to the same position of the image detector 16. At this time, the image detector 16 is a photoelectric detector and a supporting oscilloscope. Each laser pulse carries part of the information of the cell sample, and is spliced to form a time domain broadened detection image, which is finally stored in the computer 11 . Combined with the morphological changes of cells stimulated by laser at different time delays in the femtosecond scale ultrafast continuous images, it was observed that the cell membrane was dented and modified under the action of laser. Based on the time domain broadening detection image, femtosecond laser processed cells were obtained. The cell membrane produced regular perforations after 1 nanosecond, which showed that the femtosecond laser processing process was effective in perforating the cell membrane, and could be used as a pretreatment method for biological processes such as cell transfection to achieve applications such as biological detection.

Example 2

In order to detect the effects of different laser fluxes on cell activity after laser processing of cells, the device and implementation process are as follows: The pulse laser 1 generates an ultrashort pulse laser with a wavelength of 800nm and a pulse width of 50fs. The ultrashort pulse passes through a multi-pulse sequence. Generator 2 specifically generates a pulse sequence consisting of three laser pulses with wavelengths of 650nm, 800nm, and 950nm through the broadening of the nonlinear crystal and the dispersion of the optical fiber. The time delay between adjacent pulses is 300fs. The above pulse sequence is divided into two paths after passing through the first beam splitter 3. One path is focused on the cell sample 5 through the objective lens 4. The sample 5 is constrained and controlled by the microfluidic device. At this time, the laser flux irradiated on the cells is 0.5mJ/cm ² ; another laser pulse sequence acts on the sample 5 in a delayed sequence, carrying cell imaging information at different processing times and passing through four beam splitters 6-9 to spatially separate the four pulsed laser beams. The CCD camera 10 performs imaging to obtain ultra-fast continuous images of cells during laser processing at 0 fs, 300 fs, and 600 fs, and records them into the computer 11 to characterize the ultra-fast dynamic process of the cells. Next, use the cell image processing algorithm to process the above-mentioned ultra-fast continuous images of cells, perform grayscale processing, cell segmentation, cell filling, and area identification calculations. The final output of the equivalent diameter of a single cell is approximately 20 μm. At this time, the microfluidic device The cell flow rate is 10m/s. In order to obtain a detection image with a time resolution of 1μs and a time span of 2μs, the frequency of the time domain broadened pulse laser is 10 ⁸ Hz, and the speed of the microfluidic device is set to 10m/s. The time domain broadening pulse generator 13 is a combination of a high repetition frequency pulse laser and a prism, and is controlled by the cooperative controller 12. According to the above calculation results, it generates a laser with a repetition frequency of 10 ⁸ Hz and a center wavelength of 800 nm. After dispersion of the prism, The pulsed laser with a wavelength range of 790nm-810nm is separated in time, and the broadened pulsed laser is separated in a spatial position through a spatial beam splitter 14, where the spatial beam splitter is selected as a grating and acts on the flowing sample 5, where The 810nm laser reaches the sample first, and the 790nm laser reaches the sample last. The pulse laser carries the sample imaging information and enters the spatial light combiner 15 through the necessary reflectors, which converges the pulse lasers at different times and different spatial positions to the image detector. 16 At the same position, the image detector 16 is a high-speed camera, and the captured cell information is finally output and stored in the computer 11. Combined with femtosecond-scale ultrafast continuous imaging information, the stimulated morphological changes of cells in the femtosecond time scale during femtosecond laser processing of cells are obtained, as well as the multi-timescale real-time processing results that can indicate cell activity after 2 μs. The results can be judged based on the results. The cells have been damaged under the processing parameters. Adjust the parameters in time to reduce the power of pulse laser 1 to 50% of the original, so that the laser flux focused on sample 5 is 0.25mJ/cm ² . Repeat the above processing and observation process again, from Analyze information from ultra-fast continuous images and time-domain broadened detection images to achieve real-time feedback on cell processing results and optimization of processing parameters, obtain the influence of laser flux on femtosecond laser processing of cells, and support technology for high-precision non-destructive processing of cells. development, and solve related engineering and technical problems in this field.

The above-mentioned specific description further explains the purpose, technical solutions and beneficial effects of the invention in detail. It should be understood that the above-mentioned are only specific embodiments of the invention and are not intended to limit the protection of the invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection scope of the present invention.

Claims (8)

1. A multi-scale continuous observation and feedback method for femtosecond laser processing of cells, which is characterized by including the following steps: Step 1. The ultrafast laser generated by the ultrashort pulse laser is transformed into a pulse sequence with femtosecond-level time delay through a multi-frequency pulse sequence generator. The pulse sequence is divided into two beams by a beam splitting method, where One pulse sequence is focused through the objective lens and acts on the cells whose flow rate is controlled by the microfluidic device, performing multi-frequency pulsed femtosecond laser non-destructive or minimally invasive regulatory processing of the cells; after the other laser pulse sequence passes through the cells, it carries There is ultra-fast information, which is received by the CCD camera, generates ultra-fast continuous images and stores them in the computer; Step 2: Extract the cell images constrained in the microfluidic device captured by the CCD camera in Step 1. Each of the cell images contains one or more cells, and the cell equivalent diameter value d is calculated through the cell image processing method. Output to the cooperative controller; calculate the time resolution t of the time domain broadening detection image, t is the reciprocal of the laser pulse repetition frequency f of the time domain broadening pulse generator, and obtain the laser pulse repetition frequency ; According to the time span T of the required detection image and the matching relationship between the equivalent cell diameter d and the cell flow rate v : /> , calculate the cell flow rate v , set the parameters of the microfluidic device, and realize feedback adjustment of the detection process; Step 3: The cooperative controller triggers the time domain broadening pulse generator to generate the pulse laser with the repetition frequency f calculated in step 2. After the time domain broadening pulse generator generates a pulse laser with any time delay in the range of nanoseconds to milliseconds, After the laser pulse is applied to the cell, the time domain broadened detection image of the cell at this time resolution is collected by the image detector, stored in the computer, and combined with the ultra-fast continuous image received by the CCD camera to obtain a multi-scale continuous cell Observe information, observe the cell biological reaction process, and guide the optimization of femtosecond laser cell processing; the cell biological reaction process includes cell morphology, cell flow, cell membrane perforation, and cell healing. 2. The multi-scale continuous observation and feedback method of femtosecond laser processed cells as claimed in claim 1, characterized in that the implementation method of the cell image processing method in step 2 is: (1) Convert the ultra-fast continuous image containing cell morphology in step 1 into a grayscale image, and draw a grayscale histogram of the ultra-fast continuous image; (2) Select a threshold from the trough of the grayscale histogram for binary segmentation, and process the image into a binary image; (3) Fill the holes in the binary image and delete the connecting lines between cells to achieve the purpose of segmenting cell adhesions; (4) Perform morphological opening operations to improve the visual smoothness of the image; (5) Calculate the size of the coverage area of each cell and exclude block contours that are too small or too large; (6) Then find the center of gravity of the connected domain and draw numbers at the coordinate points of the center of gravity, and finally generate a cell counting marker image; (7) Output the equivalent cell diameter d. 3. A multi-scale continuous observation and feedback device for femtosecond laser-processed cells, used to implement the multi-scale continuous observation and feedback method for femtosecond laser-processed cells as claimed in claim 1 or 2, characterized in that: it includes ultra-short pulses Laser, multi-frequency pulse sequence generator, processing objective lens, microfluidic device for controlling cell flow speed, collaborative controller, time domain broadening pulse generator, spatial beam splitter, spatial light combiner, image detector, several points Beam mirror, CCD camera and computer; the ultrafast pulse laser generated by the ultrashort pulse laser is converted into a pulse sequence consisting of several multi-frequency pulse lasers with time delay through a multi-frequency pulse sequence generator, and is divided into pulse sequences through a beam splitter Two beams, one of which passes through the processing objective lens for laser processing of cells, and the other beam passes through the cells to collect cell signals within the femtosecond time scale of the cell processing process, and the multi-frequency pulse laser carrying imaging information passes through A number of beam splitters are separated, and the signals are collected by the CCD camera to generate ultra-fast continuous images and stored in the computer; the ultra-fast continuous images are identified by the cell image processing method and sent to the collaborative controller according to the steps. According to the laser pulse repetition frequency f and the cell flow rate v in the second step, the parameters of the microfluidic device are adjusted to adjust the cell flow rate, and a cooperative controller is used to trigger the time domain broadening pulse generator to generate a pulse with a nanosecond to millisecond amount. Pulse lasers of different frequencies with a time delay of 100 degrees are separated using a spatial spectrometer and act on the cells. Signals at any time scale from nanoseconds to milliseconds in the cell processing process are collected, and the spatial spectrometer is used to separate and act on the cells. The photocombiner gathers the imaging information carried by the pulse laser at different times at the same spatial position, and receives the signals through the image detector in turn to form a time-domain broadened detection image and store it in the computer; after computer processing, multi-scale cell processing imaging is obtained Information, the light stimulation and ablation process during cell processing are observed, and the cell processing process and results are fed back; the multi-scale includes femtosecond, nanosecond, microsecond or millisecond. 4. The multi-scale continuous observation and feedback device for femtosecond laser processing cells according to claim 3, wherein the ultrashort pulse laser is an ultrafast laser capable of producing a single wavelength and a pulse width of femtosecond level. . 5. The multi-scale continuous observation and feedback device for femtosecond laser processed cells according to claim 3, characterized in that: the multi-frequency pulse sequence generator includes a dispersion device, a delay device, a beam splitter, a reflector, and an objective lens. ; The dispersion device is a nonlinear crystal or optical fiber; the delay device is a quartz sheet or grating. 6. The multi-scale continuous observation and feedback device for femtosecond laser processed cells as claimed in claim 3, characterized in that: the collaborative controller utilizes the step two to calculate the cell equivalent diameter d, time domain broadening pulse generation According to the relationship between the laser pulse repetition frequency f of the device and the cell flow velocity v, set the laser pulse repetition frequency of the time domain broadening pulse generator, control the frequency of the time domain broadening laser pulse acting on the cells, and achieve the time resolution of the time domain broadening detection image. Reaching the nanosecond or millisecond level. 7. The multi-scale continuous observation and feedback device for femtosecond laser processed cells according to claim 3, wherein the time domain broadening pulse generator is a combination of a high repetition frequency pulse laser capable of generating a predetermined center wavelength and a prism. , producing a broadened laser pulse. 8. The multi-scale continuous observation and feedback device for femtosecond laser processed cells according to claim 3, wherein the image detector is a high-speed camera or a combination of a photoelectric detector and an oscilloscope depending on the time resolution. .