Drug crystal intelligent identification and sorting system and application method thereof

Abstract

The invention discloses a drug crystal intelligent identification and sorting system and an application method thereof, and the method comprises the steps: recognizing the morphology of a crystal in a microphotograph through an artificial intelligence recognition network module, and detecting whether a preset crystal droplet passes through a crystal droplet optical detection module; the single chip microcomputer judges whether the corresponding preset crystal liquid drops need to be deflected and sorted or not according to the morphology information, if yes, when the crystal liquid drop optical detection module detects that the preset crystal liquid drops pass through, the crystal liquid drop sorting circuit is controlled to be switched on, and the crystal liquid drops are sorted; a dielectrophoretic force is generated on the preset crystal liquid drops passing through the branch point of the first sorting channel and the second sorting channel, so that the preset crystal liquid drops deflect to the first sorting channel, otherwise, a crystal liquid drop sorting circuit is controlled to be switched off, and the preset crystal liquid drops directly flow into the second sorting channel; the technical effect of intelligent drug crystal sorting based on morphology difference for crystals with the same active drug components is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of drug crystal identification and sorting, and in particular to a drug crystal intelligent identification and sorting system and an application method thereof. Background Art

[0002] Crystallization is a common and economical separation technology for purifying active pharmaceutical ingredients in the pharmaceutical and chemical industries. Since the crystallization process is susceptible to a variety of physical and chemical factors (temperature, pH value, precipitant type and concentration, container size, pressure, stirring mode / speed, etc.), drug molecules of the same chemical structure may form drug crystals with different properties such as morphology and size under the same crystallization conditions. Under the same conditions of active pharmaceutical ingredients, the different morphological distributions of crystals are key indicators for measuring the quality of drug products. Traditional sensors for crystal identification are mainly based on X-ray diffraction and fluorescence labeling, but both optical sensors and chemical sensors can only recognize crystals with significant optical or chemical properties, such as protein crystals and amorphous precipitates or salt crystals. Due to the characteristics that drug crystals with the same active pharmaceutical ingredient but different morphology are difficult to detect quickly and difficult to label with specific probes, traditional optical or chemical sensors have not yet been able to effectively recognize drug crystals with different morphologies and sizes. Therefore, realizing intelligent sorting of drug crystals based on morphological differences for crystals of the same active pharmaceutical ingredient is a technical problem that needs to be solved by technicians in this field. Summary of the invention

[0003] The present invention provides a drug crystal intelligent identification and sorting system and an application method thereof, which are used to realize the intelligent sorting of drug crystals based on morphological differences of crystals with the same active drug ingredient.

[0004] In view of this, the first aspect of the present invention provides a drug crystal intelligent identification and sorting system, including a crystal droplet generation module, a microfluidic chip, a crystal droplet intelligent identification module, a single-chip microcomputer, a crystal droplet optical detection module and a crystal droplet sorting circuit;

[0005] The crystal droplet generation module is used to generate preset crystal droplets and input the preset crystal droplets into the input end of the microfluidic chip;

[0006] The microfluidic chip is used for caching the preset crystal droplets. The microfluidic chip comprises an S-shaped channel, a smooth channel and a sorting channel connected in sequence. The sorting channel comprises a first sorting channel and a second sorting channel. The first sorting channel is located above the second sorting channel.

[0007] The crystal droplet intelligent recognition module includes a microscope camera and an artificial intelligence recognition network module;

[0008] A microscope camera is used to take microscopic pictures of each pre-placed crystal droplet in the microfluidic chip;

[0009] The artificial intelligence recognition network module is used to identify the crystal morphology in each microphotograph, output the preset crystal droplet number and morphology information, and send the preset crystal droplet number and morphology information to the single chip microcomputer. The preset crystal droplet numbers are numbered in sequence according to the order in which the preset crystal droplets enter the microfluidic chip;

[0010] The crystal droplet optical detection module is used to detect whether there is a preset crystal droplet passing through the smooth channel, and send the detection result to the single chip microcomputer;

[0011] The crystal droplet sorting circuit is used to switch the circuit on and off according to the instructions of the single chip microcomputer. When the circuit is on, a dielectrophoretic force is generated on the preset crystal droplets passing through the branch point of the first sorting channel and the second sorting channel, so that the preset crystal droplets are deflected to the first sorting channel. When the circuit is off, the preset crystal droplets directly flow into the second sorting channel.

[0012] The single chip microcomputer is used to determine whether the preset crystal droplet corresponding to the preset crystal droplet number needs to be deflected and sorted according to the morphology information. If so, when the crystal droplet optical detection module detects that the preset crystal droplet passes, the crystal droplet sorting circuit is controlled to be turned on, otherwise, the crystal droplet sorting circuit is controlled to be disconnected.

[0013] Optionally, the crystal droplet optical detection module includes a light source, a baffle and a photoelectric conversion module. The light source is located above the smooth channel of the microfluidic chip, the baffle is located below the smooth channel of the microfluidic chip, a through hole is dug on the baffle, the photoelectric conversion module is located below the baffle, and the photoelectric conversion module is connected to the single-chip microcomputer.

[0014] Optionally, the crystal droplet sorting circuit includes a high-voltage power supply, a relay, a first electrode, and a second electrode, the positive electrode of the high-voltage power supply is connected to the second electrode, the negative electrode of the high-voltage power supply is connected to one end of the relay, the other end of the relay is connected to the first electrode, the relay is connected to the single-chip computer, and the first electrode and the second electrode are made of liquid metal eutectic gallium indium;

[0015] The first electrode and the second electrode are located below the branch point between the first sorting channel and the second sorting channel, and both the first electrode and the second electrode are adhered to the microfluidic chip.

[0016] Optionally, the method for manufacturing the first electrode and the second electrode includes:

[0017] Insert a metal flat needle into one end of each of the hollow first electrode and the second electrode;

[0018] Use a 1mL syringe to draw up the liquid metal eutectic gallium indium;

[0019] The first electrode, the second electrode and the microfluidic chip are placed on a heating plate and heated and cured at a heating temperature of 60°C;

[0020] On the heating plate, liquid metal eutectic gallium indium is slowly injected into the first electrode from the end of the first electrode without the metal flat needle, until the liquid metal eutectic gallium indium emerges from the top of the metal flat needle;

[0021] Then use a 1mL syringe to draw up the liquid metal eutectic gallium indium;

[0022] On the heating plate, liquid metal eutectic gallium indium is slowly injected into the second electrode from the end of the second electrode without the metal flat needle, until the liquid metal eutectic gallium indium emerges from the top of the metal flat needle;

[0023] The heating plate stops heating, and after the temperature cools down to room temperature, the excess liquid metal eutectic gallium indium is cleaned with a blade;

[0024] Use a multimeter to check the continuity of the same electrode, and use an ohmmeter to check the non-continuity of the two electrodes. There is an open circuit between the two electrodes;

[0025] The ends of the first electrode and the second electrode without the metal flat needle are sealed with insulating tape to obtain the first electrode and the second electrode made of liquid metal eutectic gallium indium.

[0026] Optionally, the first electrode and the second electrode are both barbed-hook structures.

[0027] Optionally, the crystal droplet generation module includes a silicone tube, a four-way connector, a dual-channel syringe pump, a single-channel syringe pump, a first syringe, a second syringe, and a third syringe;

[0028] The silicone tube is inserted into the four-way connector, one end of the four-way connector is connected to the S-shaped channel input end of the microfluidic chip, and the interface of the four-way connector is coated with two-component epoxy resin glue for reinforcement;

[0029] The single-channel syringe pump is connected to the first syringe, and the first syringe is connected to the opposite side of the four-way connector to the side connected to the S-shaped channel input end of the microfluidic chip;

[0030] The dual-channel syringe pump is connected to the second syringe and the third syringe, and the second syringe and the third syringe are respectively connected to two sides of the four-way connector that are not directly connected to the microfluidic chip.

[0031] Optionally, the core network of the artificial intelligence recognition network module is a DenseNet neural network containing 4 dense blocks.

[0032] Optionally, the high voltage power supply is a 3kV DC power supply.

[0033] Optionally, the through hole on the baffle has a diameter of 1 mm.

[0034] The second aspect of the present invention provides an application method of the drug crystal intelligent identification and sorting system according to any one of the first aspects,

[0035] include:

[0036] The preset crystal droplets are generated by a crystal droplet generation module, and the preset crystal droplets are transported to the S-shaped channel of the microfluidic chip at a preset flow rate;

[0037] When the preset crystal droplets are fully distributed in the S-shaped channel, the crystal droplet generation module is controlled to stop generating the preset crystal droplets;

[0038] The microscope camera of the crystal droplet intelligent recognition module sequentially takes microscopic photos of each preset crystal droplet that enters the S-shaped channel successively;

[0039] The crystal morphology in each microphotograph is identified by the artificial intelligence recognition network module of the crystal droplet intelligent recognition module, and the preset crystal droplet number and morphology information are output, and the preset crystal droplet number and morphology information are sent to the single chip microcomputer, and the preset crystal droplet numbers are numbered in sequence according to the order in which the preset crystal droplets enter the microfluidic chip;

[0040] The crystal drop optical detection module detects whether a preset crystal drop passes through, and sends the detection result to the single chip microcomputer;

[0041] The single chip microcomputer determines whether the preset crystal droplet corresponding to the preset crystal droplet number needs to be deflected and sorted based on the morphology information. If so, when the crystal droplet optical detection module detects that the preset crystal droplet has passed through, the crystal droplet sorting circuit is controlled to be turned on, and a dielectrophoretic force is generated on the preset crystal droplet passing through the branch point of the first sorting channel and the second sorting channel, so that the preset crystal droplet is deflected to the first sorting channel. Otherwise, the crystal droplet sorting circuit is controlled to be disconnected, so that the preset crystal droplet flows directly into the second sorting channel.

[0042] It can be seen from the above technical solutions that the drug crystal intelligent identification and sorting system and its application method provided by the present invention have the following advantages:

[0043] The drug crystal intelligent identification and sorting system provided by the present invention uses a microscope camera to take a microscopic photograph of each preset crystal droplet in a microfluidic chip, uses an artificial intelligence recognition network module to identify the crystal morphology in the microscopic photograph, and sends the recognition result to a single-chip microcomputer. At the same time, a crystal droplet optical detection module is used to detect whether a preset crystal droplet passes through a smooth channel. The single-chip microcomputer determines whether a preset crystal droplet corresponding to the preset crystal droplet number needs to be deflected and sorted according to the morphology information. If so, when the crystal droplet optical detection module detects that a preset crystal droplet has passed, the crystal droplet sorting circuit is controlled to be turned on, and a dielectrophoretic force is generated for the preset crystal droplet passing through the branch point of the first sorting channel and the second sorting channel, so that the preset crystal droplet is deflected to the first sorting channel. Otherwise, the crystal droplet sorting circuit is controlled to be disconnected, so that the preset crystal droplet directly flows into the second sorting channel, thereby achieving the technical effect of realizing intelligent sorting of drug crystals based on morphology differences for crystals of the same active drug ingredient.

[0044] The application method of the drug crystal intelligent identification and sorting system provided by the present invention is applied in the drug crystal intelligent identification and sorting system provided by the present invention. Its principle and technical effects are the same as those of the drug crystal intelligent identification and sorting system provided by the present invention, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0046] Figure 1 This is a schematic diagram of module connections of a drug crystal intelligent identification and sorting system provided in the present invention;

[0047] Figure 2 This is a structural principle diagram of a drug crystal intelligent identification and sorting system provided in the present invention;

[0048] Figure 3 This is a schematic diagram of a microscopic photograph of crystal droplets of the same active pharmaceutical ingredient but with different morphologies taken by a microscope camera provided in the present invention;

[0049] Figure 4 It is a schematic diagram of the droplet optical detection result of the crystal droplet optical detection module provided in the present invention;

[0050] Figure 5 The present invention provides a schematic flow chart of an application method of a drug crystal intelligent identification and sorting system. DETAILED DESCRIPTION

[0051] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings 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.

[0052] For easier understanding, see Figures 1 to 3 , the present invention provides an embodiment of a drug crystal intelligent identification and sorting system, including a crystal droplet generation module, a microfluidic chip, a crystal droplet intelligent identification module, a single-chip microcomputer, a crystal droplet optical detection module and a crystal droplet sorting circuit;

[0053] The crystal droplet generation module is used to generate preset crystal droplets and input the preset crystal droplets into the input end of the microfluidic chip;

[0054] The microfluidic chip is used for caching the preset crystal droplets. The microfluidic chip comprises an S-shaped channel, a smooth channel and a sorting channel connected in sequence. The sorting channel comprises a first sorting channel and a second sorting channel. The first sorting channel is located above the second sorting channel.

[0055] The crystal droplet intelligent recognition module includes a microscope camera and an artificial intelligence recognition network module;

[0056] A microscope camera is used to take microscopic pictures of each pre-placed crystal droplet in the microfluidic chip;

[0057] The artificial intelligence recognition network module is used to identify the crystal morphology in each microphotograph, output the preset crystal droplet number and morphology information, and send the preset crystal droplet number and morphology information to the single chip microcomputer. The preset crystal droplet numbers are numbered in sequence according to the order in which the preset crystal droplets enter the microfluidic chip;

[0058] The crystal droplet optical detection module is used to detect whether there is a preset crystal droplet passing through the smooth channel, and send the detection result to the single chip microcomputer;

[0059] The crystal droplet sorting circuit is used to switch the circuit on and off according to the instructions of the single chip microcomputer. When the circuit is on, a dielectrophoretic force is generated on the preset crystal droplets passing through the branch point of the first sorting channel and the second sorting channel, so that the preset crystal droplets are deflected to the first sorting channel. When the circuit is off, the preset crystal droplets directly flow into the second sorting channel.

[0060] The single chip microcomputer is used to determine whether the preset crystal droplet corresponding to the preset crystal droplet number needs to be deflected and sorted according to the morphology information. If so, when the crystal droplet optical detection module detects that the preset crystal droplet passes, the crystal droplet sorting circuit is controlled to be turned on, otherwise, the crystal droplet sorting circuit is controlled to be disconnected.

[0061] It should be noted that the preset crystal droplets are droplets of a supersaturated solution of indomethacin. The crystal droplet generation module includes a silicone tube, a four-way joint, a dual-channel injection pump, a single-channel injection pump, a first syringe, a second syringe and a third syringe. The silicone tube is inserted into the four-way joint. One end of the four-way joint is connected to the S-type channel input end of the microfluidic chip. The interface of the four-way joint is coated with a two-component epoxy resin glue for reinforcement. The single-channel injection pump is connected to the first syringe. The first syringe is connected to the opposite side of the four-way joint that is connected to the S-type channel input end of the microfluidic chip. The dual-channel injection pump is connected to the second syringe and the third syringe. The second syringe and the third syringe are respectively connected to the two sides of the four-way joint that are not directly connected to the microfluidic chip. Specifically, the process of generating preset crystal droplets by the crystal droplet generation module is as follows:

[0062] (1) Taking indomethacin as a representative drug, 6.5 mL of indomethacin ethanol solution (mass of indomethacin powder: 450 mg, volume of anhydrous ethanol: 15 mL) and 3.5 mL of pure water were mixed and shaken in a centrifuge tube to form an indomethacin supersaturated solution;

[0063] (2) Connect two 10 mL syringes (i.e., the second syringe and the third syringe) filled with liquid paraffin to the opposite sides of the four-way connector, and use a dual-channel syringe pump to push the two 10 mL syringes filled with liquid paraffin to ensure that the flow rate at both ends is 150 μL / min;

[0064] (3) Connecting a syringe containing a supersaturated indomethacin solution (i.e., the first syringe) to one side of the four-way connector, and using a single-channel syringe pump to push the syringe containing the supersaturated indomethacin solution, with the flow rate set to 15 μL / min;

[0065] (4) connecting one connector of the four-way connector to the input end of the S-shaped channel of the microfluidic chip to input the generated droplets into the S-shaped channel;

[0066] (5) After the generation speed of the preset crystal droplets is stable and the preset crystal droplets are fully distributed in the S-shaped channel of the microfluidic chip, the injection pump is stopped.

[0067] The production process of microfluidic chip is as follows:

[0068] 1) Use SolidWorks to draw a three-dimensional model of the chip punch. The model parameters are: the length and width of the substrate are 150×75 mm, and the width and height of the cross section of the S-shaped channel are both 1 mm;

[0069] 2) Obtaining a chip aluminum mold through mechanical processing technology;

[0070] 3) Mix the PDMS (polydimethylsiloxane) polymer and the cross-linking agent in a mass ratio of 10:1 and place in a vacuum box under negative pressure for 30 minutes;

[0071] 4) Place the chip aluminum mold with the flat surface facing downward in a square container made of tin foil, and pour PDMS into the container until the liquid level is about 5 mm higher than the highest point of the chip aluminum mold. Then put the container into a vacuum box for 30 minutes;

[0072] 5) Place the square container on a heating plate, set the heating temperature to 80°C, and the heating time to 30 minutes;

[0073] 6) using a puncher to punch holes at the inlet of the S-shaped channel and the outlet of the smooth channel integrally connected to the S-shaped channel to obtain a PDMS chip;

[0074] 7) After cleaning, place a rectangular glass plate (16.3 cm × 8.5 cm) on a horizontal surface, pour a small amount of PDMS on the surface, wait for 20 minutes until a uniform PDMS film is automatically formed, and then heat and cure at 80°C for 30 minutes;

[0075] 8) Bonding the rectangular glass plate and the PDMS chip to complete the substrate preparation of the chip;

[0076] 9) Insert a metal flat needle into the entrance of the S-shaped channel;

[0077] 10) Liquid paraffin is passed into the channel of the fabricated chip, and the chip is placed in a container filled with liquid paraffin for 24 hours.

[0078] The crystal droplet optical detection module includes a light source, a baffle and a photoelectric conversion module. The light source is located above the smooth channel of the microfluidic chip, the baffle is located below the smooth channel of the microfluidic chip, a 1mm through hole is dug on the baffle, and the photoelectric conversion module is located below the baffle. The photoelectric conversion module is connected to the single-chip microcomputer. The working principle of the crystal droplet optical detection module is as follows:

[0079] When the preset crystal droplet passes through the smooth channel position corresponding to the through hole, it will refract and reflect part of the light, causing the light received by the photoelectric conversion module to weaken. The photoelectric conversion module converts the light change into a voltage signal and sends it to the microcontroller, which recognizes the voltage signal. Figure 4 As shown, when the preset crystal droplet does not reach the smooth channel position corresponding to the through hole, the waveform image of the photoelectric conversion module has no obvious peak. When the preset crystal droplet reaches the smooth channel position corresponding to the through hole, a peak appears in the waveform image and the detected voltage signal value becomes larger.

[0080] The core network of the artificial intelligence recognition network module is a DenseNet neural network containing 4 dense blocks. The process of using the microscope camera and the artificial intelligence recognition network module to identify the crystal morphology is as follows:

[0081] 1) The microscope camera sequentially obtains microscopic photographs of the crystal droplets that enter the channel one after another along the path of the S-shaped channel, and collects multiple microscopic photographs of crystal droplets with different morphologies. The microscopic photographs of the crystal droplets can be named sequentially starting from 1;

[0082] 2) All collected crystal droplet micrographs are compiled into a data set;

[0083] 3) A 4-fold cross-validation method is used to train and evaluate the neural network. In each type of picture (microscopic pictures of crystal droplets of the same morphology), a training set is randomly selected at a ratio of 8:2, and the remaining pictures are used as the validation set. All pictures in each sub-data set used for cross-validation are divided into three folders according to categories: "without droplets", "with linear crystals" and "with rod-shaped crystals". In the embodiment of the present invention, the two morphologies of "with linear crystals" and "with rod-shaped crystals" are used as examples for explanation. In practical applications, adjustments can be made according to actual conditions.

[0084] 4) Import the DenseNet neural network parameters that have been trained on the ImageNet dataset;

[0085] 5) Use the training set to train the DenseNet neural network, lock all the bottom nodes of the DenseNet neural network, and only train the top 20 layers of neuron nodes; training parameters: the loss function is the cross entropy function with an initial learning rate of 0.0001, the optimization method is Adam, the batch size is 4, and the number of training iterations is 50.

[0086] 6) The DenseNet neural network obtained after training can recognize each image in 44ms on the validation set, and the accuracy rates of images with “no droplets”, “containing linear crystals” and “containing rod-shaped crystals” are 100%, 95% and 100% respectively.

[0087] 7) Using a microscope camera, sequentially obtain microscopic photographs of the preset crystal droplets that enter the S-shaped channel, input the microscopic photographs into the trained DenseNet neural network, obtain the crystal morphology identified by the DenseNet neural network, and output the preset crystal droplet number and morphology information.

[0088] The crystal droplet sorting circuit includes a high-voltage power supply, a relay, a first electrode and a second electrode. The positive electrode of the high-voltage power supply is connected to the second electrode, the negative electrode of the high-voltage power supply is connected to one end of the relay, the other end of the relay is connected to the first electrode, the relay is connected to the single-chip computer, the first electrode and the second electrode are made of liquid metal eutectic gallium indium, the first electrode and the second electrode are located below the branch point of the first sorting channel and the second sorting channel, and the first electrode and the second electrode are both adhered to the microfluidic chip.

[0089] The first electrode and the second electrode are both hook-shaped structures. The high voltage power supply is a 3kV DC power supply.

[0090] The method for manufacturing the first electrode and the second electrode comprises:

[0091] Insert a metal flat needle into one end of each of the hollow first electrode and the second electrode;

[0092] Use a 1mL syringe to draw up the liquid metal eutectic gallium indium;

[0093] The first electrode, the second electrode and the microfluidic chip are placed on a heating plate and heated and cured at a heating temperature of 60°C;

[0094] On the heating plate, liquid metal eutectic gallium indium is slowly injected into the first electrode from the end of the first electrode without the metal flat needle, until the liquid metal eutectic gallium indium emerges from the top of the metal flat needle;

[0095] Then use a 1mL syringe to draw up the liquid metal eutectic gallium indium;

[0096] On the heating plate, liquid metal eutectic gallium indium is slowly injected into the second electrode from the end of the second electrode without the metal flat needle, until the liquid metal eutectic gallium indium emerges from the top of the metal flat needle;

[0097] The heating plate stops heating, and after the temperature cools down to room temperature, the excess liquid metal eutectic gallium indium is cleaned with a blade;

[0098] Use a multimeter to check the continuity of the same electrode, and use an ohmmeter to check the non-continuity of the two electrodes. There is an open circuit between the two electrodes;

[0099] The ends of the first electrode and the second electrode without the metal flat needle are sealed with insulating tape to obtain the first electrode and the second electrode made of liquid metal eutectic gallium indium and adhered to the microfluidic chip.

[0100] The artificial intelligence recognition network module outputs the preset crystal droplet number and morphology information to the single-chip microcomputer, and the single-chip microcomputer determines whether the droplet needs to be deflected and sorted by the preset crystal droplet number and morphology information. If the single-chip microcomputer recognizes that the droplet contains linear crystals, the single-chip microcomputer outputs a high level to the relay, and the relay receives the high level signal, the circuit is closed, and the 3kV high-voltage power supply applies a 3kV DC voltage to the two electrodes (the first electrode and the second electrode). The droplet is deflected upward by the dielectrophoretic force under the action of the uneven electric field and flows into the first sorting channel above. If a droplet containing the required rod-shaped crystal is identified, the droplet flows into the second sorting channel below in a natural flow state.

[0101] The drug crystal intelligent identification and sorting system provided by the present invention uses a microscope camera to take a microscopic photograph of each preset crystal droplet in a microfluidic chip, uses an artificial intelligence recognition network module to identify the crystal morphology in the microscopic photograph, and sends the recognition result to a single-chip microcomputer. At the same time, a crystal droplet optical detection module is used to detect whether a preset crystal droplet passes through a smooth channel. The single-chip microcomputer determines whether a preset crystal droplet corresponding to the preset crystal droplet number needs to be deflected and sorted according to the morphology information. If so, when the crystal droplet optical detection module detects that a preset crystal droplet has passed, the crystal droplet sorting circuit is controlled to be turned on, and a dielectrophoretic force is generated for the preset crystal droplet passing through the branch point of the first sorting channel and the second sorting channel, so that the preset crystal droplet is deflected to the first sorting channel. Otherwise, the crystal droplet sorting circuit is controlled to be disconnected, so that the preset crystal droplet directly flows into the second sorting channel, thereby achieving the technical effect of realizing intelligent sorting of drug crystals based on morphology differences for crystals of the same active drug ingredient.

[0102] For easier understanding, see Figure 5 The present invention provides an embodiment of an application method of the drug crystal intelligent identification and sorting system provided in the present invention, comprising:

[0103] Step S1, generating a preset crystal droplet through a crystal droplet generation module, and delivering the preset crystal droplet to an S-shaped channel of a microfluidic chip at a preset flow rate;

[0104] Step S2, when the preset crystal droplets are fully distributed in the S-shaped channel, controlling the crystal droplet generation module to stop generating the preset crystal droplets;

[0105] Step S3, taking microscopic photos of each preset crystal droplet that successively enters the S-shaped channel in sequence through a microscope camera of the crystal droplet intelligent recognition module;

[0106] Step S4, identifying the crystal morphology in each micrograph through the artificial intelligence recognition network module of the crystal droplet intelligent recognition module, outputting the preset crystal droplet number and morphology information, and sending the preset crystal droplet number and morphology information to the single-chip microcomputer, and the preset crystal droplet numbers are numbered in sequence according to the order in which the preset crystal droplets enter the microfluidic chip;

[0107] Step S5, detecting whether a preset crystal droplet has passed through by means of a crystal droplet optical detection module, and sending the detection result to the single chip microcomputer;

[0108] Step S6, the single chip computer determines whether the preset crystal droplet corresponding to the preset crystal droplet number needs to be deflected and sorted according to the morphology information. If so, when the crystal droplet optical detection module detects that the preset crystal droplet has passed, the crystal droplet sorting circuit is controlled to be turned on, and a dielectrophoretic force is generated on the preset crystal droplet passing through the branch point of the first sorting channel and the second sorting channel, so that the preset crystal droplet is deflected to the first sorting channel. Otherwise, the crystal droplet sorting circuit is controlled to be disconnected, so that the preset crystal droplet flows directly into the second sorting channel.

[0109] The application method of the drug crystal intelligent identification and sorting system provided by the present invention is applied in the drug crystal intelligent identification and sorting system provided by the present invention. Its principle and technical effects are the same as those of the drug crystal intelligent identification and sorting system provided by the present invention, and will not be repeated here.

[0110] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A drug crystal intelligent identification and sorting system, characterized in that: It includes a crystal droplet generation module, a microfluidic chip, a crystal droplet intelligent recognition module, a single-chip microcomputer, a crystal droplet optical detection module and a crystal droplet sorting circuit; The crystal droplet generation module is used to generate preset crystal droplets and input the preset crystal droplets into the input end of the microfluidic chip; The microfluidic chip is used for caching the preset crystal droplets. The microfluidic chip comprises an S-shaped channel, a smooth channel and a sorting channel connected in sequence. The sorting channel comprises a first sorting channel and a second sorting channel. The first sorting channel is located above the second sorting channel. The crystal droplet intelligent recognition module includes a microscope camera and an artificial intelligence recognition network module; A microscope camera is used to take microscopic pictures of each pre-placed crystal droplet in the microfluidic chip; The artificial intelligence recognition network module is used to identify the crystal morphology in each microphotograph, output the preset crystal droplet number and morphology information, and send the preset crystal droplet number and morphology information to the single chip microcomputer. The preset crystal droplet numbers are numbered in sequence according to the order in which the preset crystal droplets enter the microfluidic chip; The crystal droplet optical detection module is used to detect whether there is a preset crystal droplet passing through the smooth channel, and send the detection result to the single chip microcomputer; The crystal droplet sorting circuit is used to switch the circuit on and off according to the instructions of the single chip microcomputer. When the circuit is on, a dielectrophoretic force is generated on the preset crystal droplets passing through the branch point of the first sorting channel and the second sorting channel, so that the preset crystal droplets are deflected to the first sorting channel. When the circuit is off, the preset crystal droplets directly flow into the second sorting channel. The single chip microcomputer is used to determine whether the preset crystal droplet corresponding to the preset crystal droplet number needs to be deflected and sorted according to the morphology information. If so, when the crystal droplet optical detection module detects that the preset crystal droplet passes, the crystal droplet sorting circuit is controlled to be turned on, otherwise, the crystal droplet sorting circuit is controlled to be disconnected.

2. The drug crystal intelligent identification and sorting system according to claim 1, characterized in that: The crystal droplet optical detection module includes a light source, a baffle and a photoelectric conversion module. The light source is located above the smooth channel of the microfluidic chip, the baffle is located below the smooth channel of the microfluidic chip, a through hole is dug on the baffle, the photoelectric conversion module is located below the baffle, and the photoelectric conversion module is connected to the single-chip microcomputer.

3. The drug crystal intelligent identification and sorting system according to claim 2, characterized in that: The crystal droplet sorting circuit includes a high-voltage power supply, a relay, a first electrode and a second electrode, wherein the positive electrode of the high-voltage power supply is connected to the second electrode, the negative electrode of the high-voltage power supply is connected to one end of the relay, the other end of the relay is connected to the first electrode, the relay is connected to the single-chip microcomputer, and the first electrode and the second electrode are made of liquid metal eutectic gallium indium; The first electrode and the second electrode are located below the branch point between the first sorting channel and the second sorting channel, and both the first electrode and the second electrode are adhered to the microfluidic chip.

4. The drug crystal intelligent identification and sorting system according to claim 3, characterized in that: The method for manufacturing the first electrode and the second electrode comprises: Insert a metal flat needle into one end of each of the hollow first electrode and the second electrode; Use a 1mL syringe to draw up the liquid metal eutectic gallium indium; The first electrode, the second electrode and the microfluidic chip are placed on a heating plate and heated and cured at a heating temperature of 60°C; On the heating plate, liquid metal eutectic gallium indium is slowly injected into the first electrode from the end of the first electrode without the metal flat needle, until the liquid metal eutectic gallium indium emerges from the top of the metal flat needle; Then use a 1mL syringe to draw up the liquid metal eutectic gallium indium; On the heating plate, liquid metal eutectic gallium indium is slowly injected into the second electrode from the end of the second electrode without the metal flat needle, until the liquid metal eutectic gallium indium emerges from the top of the metal flat needle; The heating plate stops heating, and after the temperature cools down to room temperature, the excess liquid metal eutectic gallium indium is cleaned with a blade; Use a multimeter to check the continuity of the same electrode, and use an ohmmeter to check the non-continuity of the two electrodes. There is an open circuit between the two electrodes; The ends of the first electrode and the second electrode without the metal flat needle are sealed with insulating tape to obtain the first electrode and the second electrode made of liquid metal eutectic gallium indium.

5. The drug crystal intelligent identification and sorting system according to claim 3, characterized in that: The first electrode and the second electrode are both in a barb-shaped structure.

6. The drug crystal intelligent identification and sorting system according to claim 1, characterized in that: The crystal droplet generation module includes a silicone tube, a four-way connector, a dual-channel syringe pump, a single-channel syringe pump, a first syringe, a second syringe and a third syringe; The silicone tube is inserted into the four-way connector, one end of the four-way connector is connected to the S-shaped channel input end of the microfluidic chip, and the interface of the four-way connector is coated with two-component epoxy resin glue for reinforcement; The single-channel syringe pump is connected to the first syringe, and the first syringe is connected to the opposite side of the four-way connector to the side connected to the S-shaped channel input end of the microfluidic chip; The dual-channel syringe pump is connected to the second syringe and the third syringe, and the second syringe and the third syringe are respectively connected to two sides of the four-way connector that are not directly connected to the microfluidic chip.

7. The drug crystal intelligent identification and sorting system according to claim 1, characterized in that: The core network of the artificial intelligence recognition network module is a DenseNet neural network containing 4 dense blocks.

8. The drug crystal intelligent identification and sorting system according to claim 3, characterized in that: The high voltage power supply is 3kV DC power supply.

9. The drug crystal intelligent identification and sorting system according to claim 1, characterized in that: The diameter of the through hole in the baffle is 1 mm.

10. An application method of the drug crystal intelligent identification and sorting system according to any one of claims 1 to 9, characterized in that: include: The preset crystal droplets are generated by a crystal droplet generation module, and the preset crystal droplets are transported to the S-shaped channel of the microfluidic chip at a preset flow rate; When the preset crystal droplets are fully distributed in the S-shaped channel, the crystal droplet generation module is controlled to stop generating the preset crystal droplets; The microscope camera of the crystal droplet intelligent recognition module sequentially takes microscopic photos of each preset crystal droplet that enters the S-shaped channel successively; The crystal morphology in each microphotograph is identified by the artificial intelligence recognition network module of the crystal droplet intelligent recognition module, and the preset crystal droplet number and morphology information are output, and the preset crystal droplet number and morphology information are sent to the single chip microcomputer, and the preset crystal droplet numbers are numbered in sequence according to the order in which the preset crystal droplets enter the microfluidic chip; The crystal drop optical detection module detects whether a preset crystal drop passes through, and sends the detection result to the single chip microcomputer; The single chip microcomputer determines whether the preset crystal droplet corresponding to the preset crystal droplet number needs to be deflected and sorted based on the morphology information. If so, when the crystal droplet optical detection module detects that the preset crystal droplet has passed through, the crystal droplet sorting circuit is controlled to be turned on, and a dielectrophoretic force is generated on the preset crystal droplet passing through the branch point of the first sorting channel and the second sorting channel, so that the preset crystal droplet is deflected to the first sorting channel. Otherwise, the crystal droplet sorting circuit is controlled to be disconnected, so that the preset crystal droplet flows directly into the second sorting channel.

Images