Multi-mode needle cylinder liquid intelligent detection method, device and system

Abstract

The invention relates to the technical field of automatic detection, and discloses a multi-mode needle cylinder liquid intelligent detection method, device and system, and the system comprises a multi-mode imaging module, a dynamic calibration module, a volume calculation module and a man-machine interaction module. The multi-modal imaging module, the dynamic calibration module, the volume calculation module and the man-machine interaction module are jointly configured with a voice alarm module; by arranging a multi-mode imaging end, when the existence of liquid in a needle cylinder is detected, dual-mode data of visible light transmission light intensity distribution and near-infrared absorption spectrum are synchronously acquired, and dual criteria of the existence of the liquid are constructed, so that the liquid level refraction characteristic and the molecular absorption characteristic can be subjected to real-time cross validation; the method guarantees the discrimination of transparent liquid, a low-contrast medium and a bubble-containing medicament, eliminates the misjudgment risk of traditional single spectral imaging, and improves the reliability of liquid existence detection.

Description

[0001] The present application relates to the field of automated detection technology, in particular to a multi-modal syringe liquid intelligent detection method, device and system. BACKGROUND

[0002] Liquid detection technology is related to various fields such as chemical production, oil storage and transportation, biomedicine, food and beverage, and water quality detection, and is one of the key technologies affecting people's livelihood, national development and ecological environment, among which, liquid level measurement is a real-time monitoring and detection process of the liquid level in a container, and liquid detection includes detection of liquid level, color and concentration characteristics, and is related to many fields of social development and technological progress.

[0003] At present, due to the factors of material diversity, complex liquid characteristics and variable space posture in the syringe liquid detection process, when detecting the existence and volume of the syringe liquid, the optical imaging system equipped cannot detect in real time whether the liquid position in the syringe is disturbed and distorted by light caused by material differences. Such light distortion will cause large liquid surface positioning error and cannot guarantee the accuracy of detection. At the same time, when measuring the volume of the liquid, the syringe space posture and taper deformation cannot be detected in real time, which will cause a volume calculation deviation of 0.4ml, and the detection system cannot correct the space modeling parameters in real time. When analyzing complex liquid media, due to the limitation of single spectrum perception, multi-spectrum fusion criterion cannot be realized, which leads to the lack of cross-validation in liquid existence determination, further affecting the reliability of detection and the requirement of medical grade precision.

[0004] Therefore, the present application proposes a multi-modal syringe liquid intelligent detection method, device and system to solve the above problems. SUMMARY

[0005] The main purpose of the present application is to provide a multi-modal syringe liquid intelligent detection method, device and system to solve the problems raised in the above background.

[0006] To achieve the above purpose, the technical solution adopted by the present application is as follows: a multi-modal syringe liquid intelligent detection method, device and system, the system comprises a multi-modal imaging module, a dynamic calibration module, a volume calculation module and a man-machine interaction module, the multi-modal imaging module, the dynamic calibration module, the volume calculation module and the man-machine interaction module are jointly configured with a voice alarm module;

[0007] The multi-modal imaging module generates a liquid existence dual criterion by synchronously collecting the visible light transmission intensity distribution and near-infrared absorption spectrum data of the syringe, and determines the liquid existence state in real time.

[0008] The dynamic calibration module identifies the needle cylinder material type through the capacitive sensor, automatically adjusts the lighting parameters of the annular array LED light source and the focal length of the automatic zoom lens group, and adapts to the characteristics of glass and plastic containers.

[0009] The volume calculation module performs three-dimensional integral operation and outputs the liquid volume value based on the needle cylinder outer diameter data and liquid level coordinate obtained by the laser triangulation range finder.

[0010] The human-computer interaction module visually displays the detection results and marks the liquid existence, container adaptability and volume over-standard abnormal state.

[0011] The voice alarm module triggers a voice alarm when detecting that the liquid existence confidence is less than 99%, the volume error is greater than 0.1ml, and the container material matching fails.

[0012] Preferably, the voice alarm module realizes abnormal response through the following steps:

[0013] Abnormal code mapping: classify the detection abnormalities into:

[0014] C1 level abnormality, mapped to the pre-recorded voice "liquid confirmation failure, please recheck";

[0015] C2 level abnormality, mapped to the voice "dose over-standard, code C2";

[0016] C3 level abnormality, mapped to the voice "container not recognized, code C3";

[0017] Acoustic-optic cooperative response:

[0018] Triggering the voice alarm synchronously lights up the annular LED warning light;

[0019] According to the medical device alarm standard, control the loudspeaker to play in a loop at 85dB sound pressure level until the manual reset;

[0020] Manual intervention record: generate an audit log containing abnormal code, trigger time and reset operation time on the human-computer interaction end, and upload it to the medical quality management platform through encrypted data chain.

[0021] Preferably, the method comprises the following steps:

[0022] S1: synchronously acquire the visible light band transmission light intensity distribution data and near-infrared band absorption spectrum data of the needle cylinder through the multi-modal imaging module;

[0023] S2: identify the liquid surface refraction feature based on the visible light band data, analyze the liquid absorption spectrum characteristics in combination with the near-infrared band data, and generate a liquid existence double criterion;

[0024] S3: dynamically adjust the lighting parameters of the annular array LED light source and the magnification of the automatic zoom lens group according to the container material type;

[0025] S4: obtain the outer diameter data of the needle cylinder by a laser triangulation range finder, combine the liquid level coordinate of S2, and calculate the actual volume of the liquid by integral method;

[0026] S5: output the liquid existence judgment result, the container adaptability state and the liquid volume measurement value, and the error range is controlled within 0.1ml.

[0027] Preferably, the multi-modal imaging module in S1 comprises:

[0028] A dual-spectrum synchronous imaging system of visible light channel and near-infrared channel;

[0029] The synchronous acquisition realizes the separation of light paths of visible light and near-infrared waveband through a light splitting prism, and the image data is captured in parallel by a high-speed CMOS sensor.

[0030] Preferably, the dynamic adjustment of S3 specifically comprises:

[0031] Recognize the material of the needle cylinder through a capacitive sensing unit;

[0032] Adjust the brightness and incident angle of the 8 independent partitions of the annular array LED light source, wherein 45° lateral illumination is adopted for glass material and 90° vertical transmission illumination is adopted for plastic material;

[0033] Control the automatic zoom lens group to continuously adjust the focal length within 5 times, and match the size of the needle cylinder.

[0034] Preferably, the volume calculation of S4 adopts a three-dimensional integral model:

[0035]

[0036] Wherein is the liquid level coordinate, is the needle cylinder outer diameter function, is the liquid volume, is the needle cylinder axial coordinate, by synchronously collecting the dual-mode data of visible light transmission intensity distribution and near-infrared absorption spectrum, a liquid existence double criterion is constructed, which can cross-verify the liquid refraction characteristics and molecular absorption characteristics in real time, and ensure the discrimination of transparent liquid, low-contrast medium and gas bubble-containing medicament.

[0037] Preferably, the liquid existence double criterion generation of S2 needs to meet the following conditions simultaneously:

[0038]

[0039] Wherein The refractive index mutation value for the visible light band, The signal-to-noise ratio of the characteristic absorption peak of liquid water for the near-infrared band, The logical AND operator.

[0040] Preferably, the system comprises:

[0041] A multi-modal imaging unit: an optical imaging assembly integrating visible light and near-infrared dual channels;

[0042] A dynamic calibration unit: containing a mechanical rotating table, a ring array LED light source, and an automatic zoom lens group;

[0043] A volume calculation unit: built-in laser triangulation and three-dimensional integral algorithm processor;

[0044] A control unit: configured with FPGA chips to realize imaging control, light source adjustment, and data fusion, to ensure the discrimination of transparent liquids, low-contrast media, and gas-containing agents, eliminate the misjudgment risk of traditional single-spectrum imaging, and improve the reliability of liquid existence detection.

[0045] Preferably, the dynamic calibration unit comprises:

[0046] A capacitive material sensor embedded in the mechanical rotating table clamp, which feeds back the dielectric constant of the container to the control unit in real time;

[0047] A real-time feedback adjustment mechanism adjusts the wavelength weight of the LED light source according to the material data.

[0048] Preferably, the laser triangulation of the volume calculation unit adopts:

[0049] Line scanning mode, obtaining the outer diameter profile with 0.1mm step along the needle cylinder axis;

[0050] The spatial coordinate alignment accuracy with the multi-modal imaging unit is less than 5μm, the capacitive sensor can real-time sense the dielectric property difference of the material, and the partition illumination angle of the ring array light source and the focal length of the automatic zoom lens group are adjusted in linkage, which can adaptively suppress the glass high reflection overexposure and plastic light transmission distortion, ensure the system to be free from manual intervention and quickly switch the detection of different material containers, combined with the liquid level height coordinate positioned by the multi-modal imaging, and using the spatial integral algorithm dynamic modeling, the needle cylinder taper deformation and liquid surface inclination error can be compensated, the dose calculation value can strictly meet the medical equipment precision standard, and the precision bottleneck in the measurement of trace liquid volume by traditional detection methods is broken through.

[0051] The present application has the following beneficial effects:

[0052] 1. In the present application, by setting the multi-modal imaging end, when the needle cylinder liquid existence detection is carried out, the visible light transmission intensity distribution and the near-infrared absorption spectrum double-mode data are collected synchronously, the liquid existence double criterion is constructed, the liquid level refraction characteristics and molecular absorption characteristics can be cross-validated in real time, the discrimination of transparent liquid, low-contrast medium and gas bubble-containing medicament is ensured, the misjudgment risk of traditional single spectrum imaging is eliminated, and the reliability of liquid existence detection is improved.

[0053] 2. In the present application, by setting the dynamic calibration end, when the needle cylinder container adaptation detection is carried out, the dielectric characteristic difference of the material is sensed in real time by the capacitive sensor, and the partition illumination angle of the annular array light source and the focal length of the automatic zoom lens group are adjusted in linkage, the glass high reflection overexposure and plastic light transmission distortion can be adaptively inhibited, the system is ensured to be free from manual intervention and quickly switches the detection of different material containers, and the problems of detection interruption and efficiency reduction caused by material difference in the traditional scheme are solved.

[0054] 3. In the present application, by setting the volume calculation end, when the medical-grade liquid volume measurement is carried out, the needle cylinder space contour data are acquired by laser three-dimensional scanning, the liquid level height coordinates are combined with the multi-modal imaging positioning, the spatial integral algorithm is adopted for dynamic modeling, the needle cylinder taper deformation and liquid level inclination error can be compensated, the dose calculation value can be ensured to strictly meet the medical equipment precision standard, and the precision bottleneck in the measurement of trace liquid volume in the traditional detection method is broken through. BRIEF DESCRIPTION OF DRAWINGS

[0055] Fig. 1 The method flowchart of the multi-modal needle cylinder liquid intelligent detection method, device and system of the present application is shown in the figure.

[0056] Fig. 2 The system flowchart of the multi-modal needle cylinder liquid intelligent detection method, device and system of the present application is shown in the figure. DETAILED DESCRIPTION

[0057] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the present application.

[0058] Please refer to Figs. 1-2 The multi-modal needle cylinder liquid intelligent detection method, device and system, the system comprises a multi-modal imaging module, a dynamic calibration module, a volume calculation module and a man-machine interaction module, and the multi-modal imaging module, the dynamic calibration module, the volume calculation module and the man-machine interaction module are collectively configured with a voice alarm module.

[0059] The multimodal imaging module generates a liquid presence double criterion by synchronously collecting the visible light transmission intensity distribution and near-infrared absorption spectrum data of the needle cylinder, and determines the liquid presence state in real time;

[0060] The dynamic calibration module identifies the material type of the needle cylinder through the capacitive sensor, automatically adjusts the lighting parameters of the annular array LED light source and the focal length of the automatic zoom lens group, and adapts to the characteristics of glass and plastic containers;

[0061] The volume calculation module performs three-dimensional integral operation and outputs the liquid volume value through the needle cylinder outer diameter data and liquid level coordinates obtained based on the laser triangulation range finder;

[0062] The human-computer interaction module visually displays the detection results, and marks the liquid presence, container adaptability and volume over-standard abnormal state;

[0063] The voice alarm module triggers a voice alarm when it detects that the liquid presence confidence is less than 99%, the volume error is greater than 0.1ml, and the container material matching fails.

[0064] The voice alarm module realizes abnormal response through the following steps:

[0065] Abnormal code mapping: classify the detection abnormalities as:

[0066] C1 level abnormality, mapped to the pre-recorded voice "liquid confirmation failure, please recheck";

[0067] C2 level abnormality, mapped to the voice "dose over-standard, code C2";

[0068] C3 level abnormality, mapped to the voice "container not recognized, code C3";

[0069] Acoustic-optic cooperative response:

[0070] Trigger the voice alarm and simultaneously light up the annular LED warning light;

[0071] According to the medical device alarm standard, control the loudspeaker to play in a loop at 85dB sound pressure level until the manual reset;

[0072] Manual intervention record: generate an audit log containing abnormal code, trigger time and reset operation time on the human-computer interaction end, and upload it to the medical quality management platform through encrypted data chain.

[0073] The system includes the following steps:

[0074] S1: Synchronously collect the visible light band transmission intensity distribution data and near-infrared band absorption spectrum data of the needle cylinder through the multimodal imaging module;

[0075] S2: Identify the liquid surface refraction characteristics based on the visible light band data, analyze the liquid absorption spectrum characteristics combined with the near-infrared band data, and generate liquid existence dual criteria;

[0076] S3: Dynamically adjust the lighting parameters of the ring array LED light source and the magnification of the automatic zoom lens group according to the container material type;

[0077] S4: Obtain the outer diameter data of the needle cylinder through the laser triangulation range finder, combine the liquid level coordinates of S2, and calculate the actual volume of the liquid by integral method;

[0078] S5: Output the liquid existence determination result, container adaptability state and liquid volume measurement value, with an error range control within 0.1ml.

[0079] The multi-modal imaging module in S1 includes:

[0080] Dual-spectrum synchronous imaging system of visible light channel and near-infrared channel;

[0081] Synchronous acquisition of visible light and near-infrared waveband light path separation through a dichroic prism, and parallel capture of image data by a high-speed CMOS sensor.

[0082] The dynamic adjustment of S3 specifically includes:

[0083] Recognize the material of the needle cylinder through the capacitance sensing unit;

[0084] Adjust the brightness and incident angle of the 8 independent partitions of the ring array LED light source, where 45° lateral illumination is used for glass material and 90° vertical transmission illumination is used for plastic material;

[0085] Control the automatic zoom lens group to continuously adjust the focal length within 5 times, matching the size of the needle cylinder.

[0086] The volume calculation of S4 uses a three-dimensional integral model:

[0087]

[0088] Wherein is the liquid level coordinate, is the needle cylinder outer diameter function, is the liquid volume, is the needle cylinder axial coordinate.

[0089] The liquid existence dual criteria generated by S2 need to meet:

[0090]

[0091] Wherein is the refractive index mutation value of the visible light band, For the near-infrared band liquid water characteristic absorption peak signal-to-noise ratio, For the logical and operator.

[0092] The system comprises:

[0093] Multi-modal imaging unit: integrated visible light and near-infrared dual-channel optical imaging components;

[0094] Dynamic calibration unit: containing mechanical rotating table, ring array LED light source and automatic zoom lens group;

[0095] Volume calculation unit: built-in laser triangulation and three-dimensional integral algorithm processor;

[0096] Control unit: configure FPGA chip to realize imaging control, light source adjustment and data fusion.

[0097] The dynamic calibration unit comprises:

[0098] Capacitive material sensor, embedded in the mechanical rotating table clamp, real-time feedback of the dielectric constant of the container to the control unit;

[0099] Real-time feedback adjustment mechanism, according to the material data to adjust the wavelength weight of the LED light source.

[0100] The laser triangulation of the volume calculation unit adopts:

[0101] Line scanning mode, along the needle cylinder axis with 0.1mm step to obtain the outer diameter profile;

[0102] The spatial coordinate alignment accuracy with the multi-modal imaging unit is less than 5μm, by combining the liquid level height coordinate of multi-modal imaging positioning, using spatial integral algorithm dynamic modeling, it can compensate the needle cylinder taper deformation and liquid surface tilt error, ensure that the dose calculation value strictly meets the medical equipment precision standard, break through the precision bottleneck of traditional detection method in micro-liquid volume measurement.

[0103] Implementation one: multi-modal imaging module liquid existence detection:

[0104] In the detection of 1ml transparent physiological saline syringe in the medical laboratory, the system first opens the visible light channel and the near-infrared channel, sets the refractive index mutation threshold to 0.35, and the liquid water absorption peak signal-to-noise ratio threshold to 12dB. Through the light splitting prism, the incident light is separated into double CMOS sensors, and the liquid surface visible light refraction image and near-infrared absorption spectrum curve are captured synchronously. When the refractive index mutation value of the liquid surface edge is identified to be 0.41 and the absorption peak signal-to-noise ratio at 1450nm is 15.2dB, the system outputs the liquid existence confirmation signal within 0.1 seconds. This process reduces the transparent liquid false detection rate from 12% of the traditional single spectrum method to 0.2%

[0105] Implementation II: Dynamic calibration module for container adaptive adjustment:

[0106] For the detection of a 5ml plastic insulin syringe, the dielectric constant measured by the capacitive sensor is 2.8, triggering illumination mode B: turning off the 4-part light source and activating the 8-part vertical transmission at 1050nm waveband, while adjusting the automatic zoom lens group to 1.2 times focal length to adapt to the plastic curvature. The system monitors the image clarity in real time to 120Lux, the material matching success rate is 100%, and the calibration time is only 0.15 seconds. When simulating glass material interference, the system automatically re-inspects and corrects the parameters within 3ms.

[0107] Implementation III: Medical-grade precision verification of volume calculation module:

[0108] In the volume detection of a 3ml plasma syringe at a blood station, the system uses a laser triangulation range finder to scan along the axial direction of the syringe with a step precision of 0.1mm, measures the conical profile data of the proximal diameter 12.05mm and the distal diameter 11.20mm, and establishes a spatial gradient model of linear decrease of diameter with axial coordinate. At the same time, the multi-modal imaging module locates the liquid level coordinate as 24.3mm. The volume calculation module performs three-dimensional integral operation based on the spatial gradient model: according to the axial gradient, it integrates the continuous cross-sectional area to calculate the actual volume occupied by the liquid, and outputs the volume value 3.02ml in real time. After calibration verification by a standard syringe, the actual volume is 3.00ml, and the system measurement deviation is only 0.02ml, which is lower than the 0.1ml error limit value required by medical equipment. To verify the performance under extreme conditions, the syringe is measured again at a 15° inclination. The traditional two-dimensional algorithm produces an error of 0.38ml due to the lack of space posture compensation, while the system compresses the error to 0.09ml through spatial integral modeling, proving its strong adaptability to complex conditions.

[0109] Implementation IV: Dynamic calibration end adaptive verification in mixed material production line:

[0110] In the scenario of simultaneously handling glass and plastic syringes in a vaccine filling production line, the system realizes automatic material switching through the dynamic calibration end: when the capacitive sensor detects a dielectric constant of 5.2 glass material, it immediately activates the 4-part LED for 45° side illumination at 850nm waveband, and adjusts the automatic zoom lens to 3 times focal length; then when switching to a plastic syringe, the system turns off the 4-part in 0.2 seconds, starts the 8-part vertical transmission at 1050nm waveband, and the focal length is adjusted to 1.5 times synchronously. During the whole process, the image clarity is stably maintained at 110Lux, and the material misjudgment rate is zero. Compared with the traditional production line solution that needs to stop for 5 minutes to replace the clamp, this system realizes continuous processing of 200 pieces per minute with high efficiency.

[0111] Implementation V: Reliability test of multi-modal imaging end for low contrast liquid:

[0112] For the detection challenge of 2ml transparent heparin sodium injection, the multimodal imaging end performs dual criterion verification: the visible light channel identifies the refractive index mutation value 0.38 of the liquid surface edge, and the near-infrared channel measures the liquid water characteristic absorption peak at 1950nm with a signal-to-noise ratio of 14.8dB. The dual criterion synchronously meets the trigger liquid existence confirmation signal, and the time consumption is 0.08 seconds. In the interference test of deliberately injecting 0.1ml bubble, the system accurately judges the liquid absence because the near-infrared signal-to-noise ratio of the bubble area is only 6.2dB, and the false detection rate tends to be zero. The traditional single visible light detection has a false detection rate of 18% under the same conditions, proving that the present technology has a key breakthrough for low-contrast liquid.

[0113] Implementation six: verification of medical-grade precision of the volume calculation end under the conditions of tilting and foreign matter interference

[0114] In the drug cold chain transportation monitoring scene, for the volume detection requirement of 5ml freeze-dried powder injection reconstituted liquid, the system starts the verification process under the double limit conditions of 35° artificial tilting and silicone oil interference. The volume calculation end scans the plastic syringe with a laser triangulation range finder at a step of 0.05mm, reconstructs a three-dimensional point cloud model, and captures the distortion and 35.2° tilting posture of the barrel, simultaneously compensates for mechanical vibration errors through an inertial measurement unit, and outputs a corrected spatial coordinate system. The multimodal imaging end cooperatively activates the visible light channel to penetrate the 0.05ml silicone oil interference layer, and identifies the true liquid surface refractive index mutation value Δn=0.33; the near-infrared channel locks the liquid surface height H=38.4mm based on the characteristic absorption peak of the liquid. The spatial integration algorithm integrates the profile function to perform volume calculation: continuous cross-sectional area integration along the axial coordinate, and outputs the volume value 5.08ml. The verification result shows that the deviation from the nominal volume of the freeze-dried powder 5.00ml is 0.08ml. Under the 35° tilting interference, the traditional laser volume instrument produces a positive error of 0.42ml, while the error of the present system is only 0.06ml; the silicone layer coverage causes a deviation of-0.38ml for the traditional detection, and the present technology maintains a precision of 0.02ml; the error is stable within the range of 0.04ml under the viscosity change at-20℃. The whole process takes 0.35 seconds, meets the 200pcs / minute production line rhythm, and passes the electronic record authentication.

[0115] Although embodiments of the present application have been shown and described, it is to be understood that various modifications, substitutions, replacements and changes can be made to these embodiments without departing from the principles and spirit of the present application, and the scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A multimodal syringe liquid intelligent detection system, characterized in that, The system includes a multimodal imaging module, a dynamic calibration module, a volume calculation module, and a human-computer interaction module. The multimodal imaging module, the dynamic calibration module, the volume calculation module, and the human-computer interaction module are all configured with a voice alarm module. The multimodal imaging module generates dual criteria for liquid presence by simultaneously acquiring visible light transmission intensity distribution and near-infrared absorption spectrum data of the syringe, and determines the liquid presence status in real time. The dynamic calibration module identifies the syringe material type through a capacitive sensor and automatically adjusts the illumination parameters of the ring array LED light source and the focal length of the automatic zoom lens group to adapt to the characteristics of glass and plastic containers. The volume calculation module performs three-dimensional integral calculations and outputs the liquid volume value based on the syringe outer diameter data and liquid level height coordinates obtained by the laser triangulation rangefinder. The human-computer interaction module displays the detection results visually and marks the presence of liquid, container compatibility, and abnormal states such as volume exceeding the standard. The voice alarm module triggers a voice alarm when it detects abnormal situations such as a liquid presence confidence level of less than 99%, a volume error greater than 0.1 ml, or a failure to match container materials.

2. The multimodal syringe liquid intelligent detection system according to claim 1, characterized in that, The voice alarm module implements anomaly response through the following steps: Exception code mapping: Classify detected exceptions as follows: A C1 level anomaly is mapped to the pre-recorded voice message "Liquid confirmation failed, please recheck". C2 level anomaly, mapped to the voice message "Dosage exceeded, code C2"; A C3 level exception, mapped to the voice message "Container not recognized, code C3"; Acoustic-optical synergistic response: When the voice alarm is triggered, the ring-shaped LED warning light will be lit simultaneously; according to the medical device alarm standard, the speaker will be controlled to broadcast the alarm in a loop at a sound pressure level of 85dB until it is manually reset by pressing a button. Manual intervention record: An audit log containing the exception code, trigger time and reset operation time is generated at the human-computer interaction terminal and uploaded to the medical quality management platform through an encrypted data chain.

3. A multimodal syringe liquid intelligent detection method, referring to the multimodal syringe liquid intelligent detection system according to any one of claims 1-2, characterized in that, The method includes the following steps: S1. Simultaneously acquire visible light band transmitted light intensity distribution data and near-infrared band absorption spectrum data of the syringe through the multimodal imaging module; S2. Based on the visible light band data, identify the refractive characteristics of the liquid surface, and combine the near-infrared band data to analyze the liquid absorption spectrum characteristics, generating dual criteria for the existence of liquid. S3. Dynamically adjust the illumination parameters of the ring array LED light source and the magnification of the automatic zoom lens group according to the container material type; S4. Obtain the outer diameter data of the syringe using a laser triangulation rangefinder, and combine it with the liquid level coordinates in S2 to calculate the actual volume of the liquid using the integration method. S5. Output the liquid presence determination result, container compatibility status, and liquid volume measurement value, with the error range controlled within 0.1ml.

4. The multimodal syringe liquid intelligent detection method according to claim 3, characterized in that, The multimodal imaging module in S1 includes: A dual-spectral simultaneous imaging system with visible and near-infrared channels; The synchronous acquisition achieves optical path separation between the visible light and near-infrared bands through a beam splitter prism, and image data is captured in parallel by a high-speed CMOS sensor.

5. The multimodal syringe liquid intelligent detection method according to claim 3, characterized in that, The dynamic adjustment of S3 specifically includes: The syringe material is identified by a capacitive sensing unit; the brightness and incident angle of the eight independent zones of the ring array LED light source are adjusted, with 45° side illumination for glass materials and 90° vertical transmission illumination for plastic materials; the automatic zoom lens group is controlled to continuously adjust the focal length within a 5x range to match the syringe size.

6. The multimodal syringe liquid intelligent detection method according to claim 3, characterized in that, The volume calculation of S4 uses a three-dimensional integral model: Where H is the liquid level height coordinate, D(z) is the syringe outer diameter function, V is the liquid volume, and z is the syringe axial coordinate.

7. The multimodal syringe liquid intelligent detection method according to claim 3, characterized in that, The dual criteria for the existence of liquid in S2 must be satisfied simultaneously: (R vis >0.3)∧(SNR NIR ≥10dB) Where R vis SNR is the abrupt change in refractive index in the visible light band. NIR is the signal-to-noise ratio of the characteristic absorption peak of liquid water in the near-infrared band, and ∧ is the logical AND operator.

8. A multimodal syringe liquid intelligent detection device, comprising the multimodal syringe liquid intelligent detection method according to any one of claims 3-7, characterized in that, The device includes: An optical imaging component integrating visible light and near-infrared dual channels; Dynamic calibration unit: includes a mechanical rotary table, a ring array of LED light sources, and an automatic zoom lens assembly; Volume calculation unit: Built-in laser triangulation rangefinder and three-dimensional integration algorithm processor; Control unit: Configurable with FPGA chip to realize imaging control, light source adjustment and data fusion.

9. The multimodal syringe liquid intelligent detection device according to claim 8, characterized in that, The dynamic calibration unit includes: A capacitive material sensor is embedded in a mechanical rotary table fixture to provide real-time feedback of the container's dielectric constant to the control unit; a real-time feedback adjustment mechanism adjusts the wavelength weight of the LED light source based on the material data.

10. The multimodal syringe liquid intelligent detection device according to claim 8, characterized in that, The laser triangulation rangefinder of the volume calculation unit adopts a line scanning mode, acquiring the outer diameter profile in 0.1mm steps along the cylinder axis; the spatial coordinate alignment accuracy with the multimodal imaging unit is less than 5μm.

Images