Green onion processing and cleaning method combining visual identification and remote monitoring

Abstract

The present application relates to the technical field of agricultural product processing, and discloses a scallion processing and cleaning method combining visual identification and remote monitoring, which comprises the following steps: through synchronous acquisition of visible light and 950nm near-infrared band reflection signals and voltage comparator processing, the optical characteristics of organic stains are directly mapped into logic level signals, and based on logic judgment, a high-voltage or low-voltage cleaning mechanism is driven to perform differential flushing, the essence of stain composition is identified and targeted cleaning is realized, the cleaning energy is adaptively adjusted through the fluctuation characteristics of the reflection signals, the decontamination effect is ensured while physical damage is avoided, and the water flow pressure energy is converted into microscopic cleaning force by using a vortex cavitation nozzle, so that the cleaning depth is significantly improved.

Description

TECHNICAL FIELD

[0001] The present application relates to a scallion processing and cleaning method combining visual recognition and remote monitoring, belonging to the technical field of agricultural product processing. BACKGROUND

[0002] In the field of automatic processing of agricultural products, the cleaning technology based on optical recognition has long relied on visible light image analysis to identify surface stains. The existing mainstream solution usually uses an RGB camera to capture the color and texture characteristics of the object, and compares them with a preset model through an algorithm to determine the stains. However, this technical path has a fundamental bottleneck: color characteristics cannot represent the nature of the material composition, resulting in a high similarity in optical characteristics between organic stains such as muddy soil containing water and the outer layer of scallion skin, forming an irreconcilable recognition contradiction.

[0003] In typical scallion processing scenarios, the above contradiction directly leads to systematic failure: when mixed stains pass through the detection area, the system cannot distinguish between similar optical characteristics of foreign objects and the body, and its decision logic is forced to rely on empirical thresholds, which results in mechanical damage caused by the blind action of high-pressure water flow on the skin area, or the inability to remove deep organic attachments through low-pressure flushing, forming a vicious cycle of ineffective cleaning and excessive damage. To compensate for the recognition defects, existing technologies generally use strategies such as increasing the flushing time or using high-pressure flushing globally, which not only causes water resource waste to be higher than the industry average, but also fails to avoid the risk of hidden corruption caused by microstructure damage due to the lack of a perception mechanism for the biomechanical response.

[0004] Although the industry has tried to introduce multispectral imaging or high-precision sensors to improve recognition rates, these solutions require complex image processing algorithms and central controllers, which significantly increase hardware costs and maintenance difficulties. The millisecond-level decision delay further causes the response of the execution mechanism to lag, making it impossible to adapt to the pace of high-speed production lines. The deeper core contradiction lies in the fact that existing technologies separate material composition recognition and execution control into two independent subsystems, failing to establish a direct, real-time, and reliable physical-level mapping mechanism from optical characteristics to mechanical actions. Therefore, how to break through the essential limitations of color recognition through a physical-level signal conversion mechanism and simultaneously achieve the essential identification of stain composition, real-time perception of mechanical response, and precise control of energy release during the conveying process has become a technical problem to be solved by the present application. SUMMARY

[0005] The present application provides a scallion processing and cleaning method combining visual recognition and remote monitoring, which aims to solve the problems of inaccurate cleaning strategy, resource waste, and damage to agricultural products caused by the inability of existing optical recognition technology to distinguish the nature of material composition.

[0006] In order to achieve the above object, the application provides a green onion processing and cleaning method combining visual identification and remote monitoring, which comprises the following steps.

[0007] In step a, the reflection intensity signal of the green onion surface to the visible light band is obtained on the conveying path of the green onion; meanwhile, the reflection intensity signal of the green onion surface to the near-infrared band with a center wavelength of 950 nm is obtained.

[0008] In step b, the reflection intensity signal of the near-infrared band is input to a first voltage comparator, and when the reflection intensity signal of the near-infrared band is lower than the organic matter light absorption threshold value determined by pre-calibration, the first voltage comparator outputs a first logic high level signal.

[0009] In step c, the reflection intensity signal of the visible light band is input to a second voltage comparator, and when the reflection intensity signal of the visible light band is lower than the green onion existence threshold value determined by pre-calibration, the second voltage comparator outputs a second logic high level signal.

[0010] In step d, based on the AND logic judgment of the first logic high level signal and the second logic high level signal, a high-pressure cleaning mechanism is driven to perform strong washing on the green onion, and the strong washing is performed when the first logic high level signal is true and the second logic high level signal is true.

[0011] In step e, when the first logic high level signal is false and the second logic high level signal is true, a low-pressure cleaning mechanism is driven to perform regular washing on the green onion.

[0012] Preferably, the reflection intensity signal of the visible light band is obtained by a first light-sensitive sensing unit, and the reflection intensity signal of the near-infrared band with a center wavelength of 950 nm is obtained by a second light-sensitive sensing unit, and the second light-sensitive sensing unit is only sensitive to the near-infrared light with a center wavelength of 950 nm.

[0013] Preferably, the high-pressure cleaning mechanism and the low-pressure cleaning mechanism are both independent water paths controlled by electromagnetic valves, and the first logic high level signal and the second logic high level signal directly control the opening and closing of the electromagnetic valves through a logic gate circuit.

[0014] Preferably, the step of strong washing further comprises the following steps: obtaining the reflection intensity signal of the second light-sensitive sensing unit at the same time when the high-pressure cleaning mechanism starts washing; extracting the alternating current fluctuation component of the reflection intensity signal through a high-pass filter; inputting the alternating current fluctuation component to a third voltage comparator, and when the alternating current fluctuation component exceeds the vibration amplitude threshold value, adjusting the energy release mode of the high-pressure cleaning mechanism based on the alternating current fluctuation component; the energy release mode is to adjust the single cleaning action of the high-pressure cleaning mechanism to a gradient cleaning sequence composed of at least two pressure-increasing or time-increasing pulses.

[0015] Preferably, the adjustment condition of the energy release mode is that the difference between the AC fluctuation component of the alternating current and the AC fluctuation component of the reflection intensity signal of the first photosensitive sensing unit exceeds the differential fluctuation threshold value determined by pre-calibration.

[0016] Preferably, the end of the high-pressure cleaning mechanism is provided with a vortex cavitation nozzle, and the vortex cavitation nozzle has an internal channel structure that reduces the local pressure of the high-pressure water flow passing through the internal channel structure to below the saturated vapor pressure of water, thereby forming cavitation micro-jets on the surface of the scallion.

[0017] Preferably, the internal channel structure of the vortex cavitation nozzle includes a spiral tapered channel that increases the water flow velocity, thereby causing the local pressure to decrease, satisfying the Bernoulli principle, that is: , wherein, is the local pressure of the water flow, is the local flow rate of the water flow, is the density of the water flow, is the energy constant in the Bernoulli principle.

[0018] Preferably, the water pressure of the high-pressure cleaning mechanism is higher than 1.5 MPa.

[0019] Preferably, the method further comprises the following steps: after a predetermined time delay after each grading cleaning action is performed, the reflection intensity signal of the near-infrared waveband with a center wavelength of 950 nm of the second photosensitive sensing unit is acquired again; and the reflection intensity signal acquired again is input to the fourth voltage comparator, and when the signal is lower than the water film residue detection threshold value determined by calibration on the healthy scallion skin, the fourth voltage comparator outputs a rejection instruction, the rejection instruction is used to indicate that the scallion has quality defects, and drives the rejection device to reject the scallion from the production line.

[0020] Preferably, the range of the predetermined time delay is 30 ms to 80 ms.

[0021] Compared with the prior art, the beneficial effects of the present application are:

[0022] 1. By using independent acquisition of visible light and 950 nm near-infrared waveband reflection signals, the voltage comparator directly converts the optical characteristics into logic level signals, which bypasses the traditional image processing process, and converts the characteristic absorption of organic stains to near-infrared from the physical layer into electrical instructions that can drive the execution mechanism. This direct mapping based on the optical nature of matter distinguishes the system from traditional color recognition modes and achieves essential differentiation between organic stains and agricultural products on the conveying path.

[0023] 2. When the high-pressure cleaning is triggered, the AC fluctuation component of the near-infrared reflection signal is extracted synchronously, and the fluctuation characteristics are implicitly related to the mechanical response of the onion structure under water flow impact: when the fluctuation amplitude exceeds the preset threshold, it indicates that the target structure has a fragile risk, and the system automatically converts the single cleaning action into a gradient pressure sequence, through the progressive energy release mode of pre-wetting to loosen the stains and then main impact, while maintaining the decontamination efficiency, avoiding irreversible damage to the microscopic cell structure.

[0024] 3. When the high-pressure water flow passes through the nozzle with a spiral tapered channel, a local low-pressure area is generated according to Bernoulli's principle, and a cavitation bubble group is spontaneously formed. When the cavitation bubble collapses on the surface of the agricultural product, the micro-jet released in the microscopic scale overcomes the van der Waals force, and the micron-sized particles embedded in the epidermal wrinkles are stripped. This process does not rely on additional energy or control units, but only through the geometric design of the flow channel to convert the pressure energy of the macroscopic water flow into the microscopic interfacial cleaning force, achieving a substantial improvement in cleaning depth.

[0025] 4. At a fixed delay time after each cleaning action, the thickness difference of the residual water film is detected by the 950nm band reflection intensity, and the putrefactive tissue maintains a thicker water film due to the enhanced hydrophilicity, resulting in a near-infrared signal that remains below the threshold of healthy tissue. This mechanism reuses the physical traces generated during the cleaning process and the existing optical sensing unit, without adding hardware, to expand the simple cleaning action into an early screening link for quality defects, achieving functional coupling of the processing link and quality control. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 is a comparison chart of the near-infrared reflection signal voltage decay curve of the present application over time;

[0027] Figure 2 is a timing chart of the delay detection and rejection process after cleaning of the present application;

[0028] Figure 3 is a curve graph of water pressure change over time under different cleaning modes of the present application;

[0029] Figure 4 is a vibration amplitude judgment flow chart of the adaptive cleaning mode switching of the present application.

[0030] The purpose of the present application, functional characteristics and advantages will be further described with reference to the accompanying drawings. DETAILED DESCRIPTION

[0031] In order to make the purpose, technical solutions and advantages of the present application clearer, the specific embodiments of the present application will be described in detail below, but it should be understood that the specific embodiments described herein are intended to explain the present application, not to limit the present application.

[0032] The method for processing and cleaning green onions by fusing visual identification and remote monitoring provided by the embodiment of the application systematically integrates optical sensing based on the nature of material composition, real-time logic control at the physical level, adaptive energy regulation based on mechanical response, microscopic cleaning based on fluid dynamics optimization, and quality defect invasive detection based on physical traces after cleaning, and the like multiple stages, to jointly construct a logically closed loop and efficient and automated processing and quality control process; the specific engineering parameter settings of the hardware system of the application are as follows, wherein the spiral tapered channel of the eddy current cavitation nozzle adopts a structure with an inlet inner diameter of 5.0 mm, an outlet inner diameter of 1.5 mm, an effective channel length of 25 mm, and an internal spiral lead of 15 mm; the cut-off frequency of the high-pass filter is set to 40 Hz, which is between the noise frequency range and the signal main peak frequency range; and the working water pressure of the high-pressure cleaning mechanism is set in the interval of 1.4 MPa to 1.8 MPa, wherein the single-pulse high-pressure flushing reference pressure and the main impact pressure of the gradient cleaning sequence in the embodiment are both uniformly 1.5 MPa. The cut-off frequency is set to 40 Hz, which is between the noise frequency range and the signal main peak frequency range; the working water pressure of the high-pressure cleaning mechanism is set in the interval of 1.4 MPa to 1.8 MPa, wherein the single-pulse high-pressure flushing reference pressure and the main impact pressure of the gradient cleaning sequence in the embodiment are both uniformly 1.5 MPa. The cut-off frequency is set to 40 Hz, which is between the noise frequency range and the signal main peak frequency range; the working water pressure of the high-pressure cleaning mechanism is set in the interval of 1.4 MPa to 1.8 MPa, wherein the single-pulse high-pressure flushing reference pressure and the main impact pressure of the gradient cleaning sequence in the embodiment are both uniformly 1.5 MPa.

[0033] In the production line deployment scenario, the green onions enter the detection and cleaning area with the conveying mechanism, and the core technical obstacle faced here is that the traditional visible light imaging cannot distinguish the water-containing organic stains attached to the surface of the green onions and the dry or wrinkled epidermis of the green onions in terms of physical nature. The high optical similarity of the two in the visible light band is the root cause of the identification blind area and strategy misalignment of the existing technology. To overcome this challenge, the procedure adopted by the present application is set to collect optical reflection signals of two different wave bands in parallel on the conveying path. Specifically, the reflection intensity signal of the green onion surface to the visible light band is obtained through a first photosensitive sensing unit, and at the same time, the reflection intensity signal of the same area to the specific near-infrared band is obtained through a second photosensitive sensing unit with high sensitivity only to near-infrared light with a center wavelength of 950 nanometers. The physical basis of this design is that the stains rich in organic matter have a strong characteristic absorption effect on near-infrared light with a center wavelength of 950 nanometers, while the healthy plant tissue of the green onion shows high reflectivity. This inherent difference derived from the spectral characteristics of the material composition provides the system with an indisputable basis for judgment beyond the appearance color and texture, thereby realizing the essential identification of organic pollutants at the physical level. In view of the stringent real-time decision-making requirements of high-speed automatic production lines, the inherent millisecond-level processing delay of the central processing unit in the traditional scheme for image comparison and analysis constitutes an insurmountable performance bottleneck. The present application avoids this by a direct electrical logic processing procedure, which directly inputs the two aforementioned reflection intensity signals into independent voltage comparators for physical-level instant processing. The reflection intensity signal of the near-infrared band is fed into the first voltage comparator, and the comparison reference inside is a pre-calibrated organic matter light absorption threshold. The calibration process of this threshold is designed as follows: under the standard production line lighting environment, repeatedly measure various typical organic stain samples using the second photosensitive sensing unit, and take the statistical average of the stable output voltage signal as the threshold. When the real-time measured near-infrared reflection intensity signal is lower than this threshold due to absorption by the stain, it indicates that the target organic matter is detected, and the first voltage comparator outputs a first logic high signal. At the same time, the reflection intensity signal of the visible light band is fed into the second voltage comparator, and the comparison reference inside is a green onion presence threshold. The calibration process of this threshold is to measure the background reflection intensity under the no-load running state of the conveying belt, and record the reflection intensity of the standard specification green onion when it passes through. Take a stable intermediate value between the two as the threshold. Therefore, when the real-time signal is lower than this threshold, it indicates that the green onion body is passing through the detection area, and the second voltage comparator outputs a second logic high signal.

[0034] Further, the two logic level signals are directly sent to a hardware logic gate circuit for AND logic judgment, and the output level of the judgment result is directly used to control two independent water paths controlled by electromagnetic valves, which constitute a high-pressure cleaning mechanism and a low-pressure cleaning mechanism, respectively; when the first logic high level signal and the second logic high level signal are both true, that is, the system determines that there are scallions and organic stains attached to them in the current detection area, and the output of the logic judgment circuit drives the electromagnetic valve of the high-pressure cleaning mechanism to open instantaneously, and performs strong flushing on the area; correspondingly, when the first logic high level signal is false and the second logic high level signal is true, that is, it is determined that there are scallions but no organic stains are found, the electromagnetic valve of the low-pressure cleaning mechanism is driven to open, and normal low-pressure flushing is performed; this architecture builds a physical level mapping path from the optical characteristics to the mechanical execution mechanism without software delay, ensuring the high synchronization of the triggering of the cleaning action and the accurate positioning of the stain in time and space, however, the strong flushing process itself also accompanies the potential risk of causing physical damage to agricultural products, especially for some scallions with relatively fragile texture, the constant high-pressure water flow may cause damage to the epidermis and even the internal tissue microstructure, in order to avoid such risks, the application also embeds an adaptive energy adjustment mechanism in the strong flushing step; at the moment when the high-pressure cleaning mechanism is controlled to start flushing, the system will simultaneously start a high-pass filter to extract the alternating fluctuation component from the output signal of the second photosensitive sensing unit in real time, the amplitude of the alternating fluctuation component is directly related to the mechanical vibration response amplitude of the scallion body under water flow impact; the alternating fluctuation component is then input to the third voltage comparator for uninterrupted comparison with a preset vibration amplitude threshold, the threshold setting procedure is designed as an offline experiment process: select multiple batches of scallion samples, apply continuous increasing water pressure impact from low to high, and use high magnification microscopic equipment to synchronously observe the integrity of the epidermal cell structure, and strictly set the reflected signal fluctuation amplitude corresponding to the critical impact force at which irreversible micro-damage begins as the vibration amplitude threshold; in actual operation, once the real-time monitored alternating fluctuation component exceeds this threshold, it indicates that the current impact energy may cause damage to the target scallion body, and the system will adjust the original single high-pressure cleaning action to a gradient cleaning sequence composed of at least two pressure increasing or time length increasing pulses based on this judgment, for example, by first performing a low-pressure pre-wetting pulse to loosen the stain, and then applying a progressive energy release mode of the main impact pressure pulse, so as to ensure the stain removal efficiency while realizing intelligent avoidance of physical damage. The vibration amplitude threshold and the reference time length The setting is completed through a set of joint offline calibration procedures, which first selects the most fragile sample of the onion body in the test bench with the same light and electrical environment as the production line, and uses a stepping motor driven pressure controller to apply a series of water pressure pulses from 0.5 MPa with a step of 0.05 MPa. The alternating current fluctuation component of the second photosensitive sensing unit is synchronously collected during the action of each pulse , and the sample impact area is observed through a 100 times microscope immediately after the pulse ends. The first appearance of more than 5 cell wall ruptures or permanent deformations in a 1 square millimeter field of view is defined as microscopic damage, and the statistical average of the peak value corresponding to the last pressure level before the critical pressure is recorded . After multiplying this average value by a safety factor of 0.9, the determined is determined as the critical pressure . Then, healthy onion samples with standard adhesive organic stains are selected, and single pulse washing with a standard high pressure is applied with a pulse length of 20 ms, with a step of 10 ms. The residual rate of the stain after washing is measured by weighing method, and the shortest washing time corresponding to the first time the residual rate is less than 1% is determined as the main impact pulse length in the gradient cleaning sequence, that is, the reference length .

[0035] To further improve the peeling ability of micro-attached particles, the spray end of the high-pressure cleaning mechanism is specially configured with a vortex cavitation nozzle designed according to the principle of fluid dynamics. Its core structure is an internal spiral tapered channel. When the pressure is stably maintained at a high-pressure water flow of more than 1.5 MPa flowing through the specially designed channel, according to the fluid energy conservation relationship described by Bernoulli's principle, i.e. , where is the local pressure of the water flow, is the local flow rate of the water flow, is the water flow density, is the energy constant, and the sharp increase in water flow velocity will cause the local pressure of the water flow to Correspondingly, it suddenly drops below the saturated vapor pressure of water; this purely physical process causes the spontaneous and intensive formation of microcavitation bubbles in the water flow, which collapse in an instant when they hit the surface of the green onion and release energy-concentrated cavitation microjets within microseconds, which can produce a few atmospheres of transient impact and strong shear effect on a microscopic scale, enough to overcome the van der Waals force that binds micron-sized dust particles to the deep folds of the epidermis, thus achieving a depth of cleaning that conventional water jet technology cannot achieve. Ultimately, the invention also ingeniously reuses the physical traces generated by the cleaning action itself to seamlessly integrate the online quality control function into the end of the processing flow; Specifically, after each grading cleaning action is completed and a predetermined time delay of between thirty milliseconds and eighty milliseconds has elapsed, the system will again drive the second photosensitive sensing unit to obtain the reflection intensity signal of the near-infrared band with a central wavelength of 950 nanometers at that moment. The setting of the predetermined time delay is based on the physical property difference between the hydrophobicity of healthy onion epidermis and the hydrophilicity of damaged or initially corrupted tissue; during this delay period, the residual water film on the surface of healthy tissue will quickly drain or significantly thin out, while damaged or corrupted tissue will maintain a relatively thicker water film due to its increased hydrophilicity, and the thicker water film will produce stronger absorption of near-infrared light, resulting in a significant decrease in reflection signal intensity; input this re-acquired reflection intensity signal into the fourth voltage comparator and compare it with a pre-calibrated water film residual detection threshold; the threshold is calibrated by statistically analyzing the near-infrared reflection intensity of a large number of clean and healthy onion samples after cleaning and after the above-mentioned standard delay, thus establishing a signal baseline representing the healthy state; when the real-time signal intensity is lower than this threshold, it indicates that abnormal water film residue has been detected, and the fourth voltage comparator will output a rejection instruction to drive the rejection device at the end of the production line to accurately separate the green onions that have been judged to have hidden quality defects from the main stream of products; The dynamic updating mechanism of the stain optical feature library has a background analysis program that specifically integrates a DBSCAN clustering algorithm based on density, with the neighborhood radius parameter Eps set to twice the standard deviation of all near-infrared reflection intensity signal values in the existing feature library, and the minimum point parameter MinPts fixed at 100; During system operation, all near-infrared signal values that trigger high-pressure cleaning events are input into the algorithm model in real time, and when the number of noise points that are not part of any known cluster is more than 500 within a continuous one-hour monitoring period, the system determines that a new data cluster with significant statistical differences from the existing feature library has been detected, and automatically triggers an alarm on the remote monitoring terminal, while highlighting the timestamp and production line location information corresponding to this batch of noise points to guide physical sampling and subsequent feature library updating operations.

[0036] At the same time, when the high-pressure water flow impacts and effectively removes the organic stains on the surface of the green onion, due to the strong absorption characteristics of the stain debris to the 950nm near-infrared light, the AC fluctuation component of the second photosensor unit It will produce violent and random fluctuations, and the visible light AC fluctuation component of the first photosensitive sensor unit It is not sensitive to this, resulting in the difference signal between the two On the contrary, when the impact energy mainly acts on the onion itself and causes damage risk, the signal fluctuations of the two channels are derived from the macroscopic physical vibration of the onion, and their waveforms are highly correlated, and the difference is maintained at a low level; therefore, the differential fluctuation threshold The calibration procedure is set as follows: first, standard high-pressure flushing is applied to group B samples, and the difference signal is recorded and calculated simultaneously. Peak collection , and then take the statistical mean of the set , and set the threshold to In actual operation, only when and Only when these two conditions are met at the same time will the system finally determine that the current impact poses a risk of structural damage and switch to the gradient cleaning sequence, thereby effectively distinguishing the violent signal generated by stain removal from the onion body overload vibration.

[0037] Baseline duration in gradient wash sequence Water film residual detection threshold in quality inspection The setting is completed through a set of joint calibration procedures. The procedure first applies a single pulse high voltage of 1.5 MPa to the samples in group B. The pulse duration starts from 20 milliseconds and increases in steps of 5 milliseconds. The stain residual rate after each rinse is measured by weighing method. The shortest rinse time corresponding to the first residual rate below 1% is determined as the benchmark time. ; Then, 50 samples from group H and group D were taken, and the obtained attenuation curve was used to find the optimal function under the same wetting conditions. Calculate the optimal detection delay time with maximum signal separation , and the water film residual detection threshold Set in At the arithmetic mean of the two groups of sample signal voltages at the moment, that is, This method integrates the originally independent parameter settings into a data-driven optimization process based on measurable physical endpoints.

[0038] Embodiment 1: This embodiment is in a green onion processing center with a high-throughput processing, and its production line faces a batch of green onions just harvested from the wet soil after the rain. The batch of raw materials presents a very challenging complex working condition, which not only generally has high-moisture and sticky organic soil dirt attached, but also has significant differences in size, texture and maturity, and mixed with some tender green onions with fragile texture. The traditional cleaning strategy relying on global high pressure or fixed time length will inevitably lead to a technical dilemma in this scenario, that is, strong washing will damage the microstructure of tender green onions, and mild washing cannot effectively remove sticky soil dirt, resulting in the negative results of incomplete cleaning and physical damage.

[0039] When a green onion with wet soil dirt and dry dust attached to the surface enters the detection area, the system acquires visible light and near-infrared band reflection signals with a center wavelength of 950 nanometers in parallel, and performs physical-level voltage comparison and logical judgment. Given the strong absorption of wet organic soil dirt to near-infrared light, the signal strength of the second photosensitive sensing unit will decrease sharply, triggering the first logic high-level signal, while the first photosensitive sensing unit confirms the presence of green onions and triggers the second logic high-level signal. The AND logic judgment result immediately drives the high-pressure cleaning mechanism to start. Here, the system first resolves the fundamental contradiction between cleaning intensity and accuracy. It does not blindly perform washing, but completely entrusts the trigger right of high-pressure water flow to the organic stain recognition result based on the spectral characteristics analysis of the material composition, thereby ensuring the targeted release of energy. In other words, when the high-pressure water flow hits a tender green onion with relatively fragile texture, another layer of synergistic mechanism of the scheme is activated. The mechanical vibration of the green onion body caused by the water flow impact is captured by the second photosensitive sensing unit as an alternating fluctuation component of the reflection signal, and the size of the alternating fluctuation component directly represents the stress state of the green onion body. Through the analysis of this fluctuation component by the high-pass filter and the third voltage comparator, once its amplitude exceeds the preset vibration amplitude threshold, the system determines that the current impact energy has a damage risk, and immediately adjusts the subsequent single high-pressure cleaning action to a gradient cleaning sequence composed of low-pressure pre-wetting and high-pressure main impact. Here, the adaptive energy adjustment mechanism based on mechanical response forms a key synergistic effect with the front-end stain recognition mechanism. The former takes the judgment of the latter as the execution premise, and the latter relies on the dynamic adjustment of the former to avoid application limitations. Both of them together ensure that when the system faces complex and variable cleaning objects, its operation mode can achieve a certain balance driven by internal logic between the upper limit of decontamination effect and the bottom line of avoiding physical damage.

[0040] More deeply, the technical boundary of the cleaning process is redefined by reusing the physical traces in the processing flow. In the traditional cognition, the residual water film produced after cleaning is useless and needs to be removed. However, the present scheme converts it into a key information carrier for quality detection. When a seemingly clean scallion with a deeply cleaned surface by the vortex cavitation jet passes through the re-inspection area in the later stage, the system re-acquires the reflection signal of the 950-nanometer near-infrared band after a predetermined time delay of 30 milliseconds to 80 milliseconds. If the scallion has hidden defects such as internal bruising or early corruption, the hydrophilicity of the damaged tissue will cause the local water film to thicken, thereby causing the near-infrared reflection signal to be lower than the water film residual detection threshold calibrated by healthy samples, and finally triggering the rejection instruction. This procedure does not resort to adding independent non-destructive defect detection equipment, but reinterprets the physical phenomena left by the cleaning action, so that the cleaning process itself derives the function of quality control. This is an embodiment of the problem redefinition design idea. The original difficult online quality screening problem is transformed into a low-cost and high-efficiency signal comparison task under the new framework.

[0041] Example 2: To objectively verify the actual performance and engineering implementability of the core technical features in the present scheme, especially to quantitatively evaluate the comprehensive ability of realizing differentiated precise cleaning and self-adaptive damage avoidance under complex working conditions, this verification test is specially set up and executed.

[0042] The test is carried out on the platform of an industrial production line, which is integrated with a variable-speed controlled conveying system, above which an optical detection module containing a first photosensitive sensing unit and a second photosensitive sensing unit is deployed, a signal processing and control unit based on the direct electrical logic program of voltage comparator and hardware logic gate circuit, and a high-pressure cleaning mechanism and a low-pressure cleaning mechanism respectively configured with a vortex cavitation nozzle are cascaded controlled; in order to ensure the effectiveness and comparability of the test data, the test samples are pre-set as three groups, group A is healthy green onions with dry floating soil attached, group B is healthy green onions smeared with a calibrated high-viscosity organic dirt, and group C is green onions with a relatively fragile texture but also smeared with the high-viscosity organic dirt, which are respectively used to simulate three typical working conditions of light pollution, heavy pollution and heavy pollution and easy to break, before the test starts, the core control parameters are precisely calibrated, wherein the setting of the vibration amplitude threshold value relied on by the adaptive energy regulation mechanism aims to establish a deterministic decision boundary to distinguish between normal cleaning impact and damage impact, the calibration follows the following operation procedures: select a batch of standard C group samples, apply gradient pressure pulses from low to high with a step size of zero point one megapascal, during the action of each pulse, the peak value of the alternating fluctuation component in the near-infrared reflection intensity signal output by the second photosensitive sensing unit is recorded synchronously by the data acquisition system, and immediately after the end of each pulse, the surface microstructure of the sample impact area is checked by a high-resolution digital microscope, the peak value of the alternating fluctuation component measured at the previous pressure level corresponding to the critical pressure at which the first observation of the cell wall appears to be slightly damaged is set as the vibration amplitude threshold value, for the sample characteristics used in this test, the threshold value is determined to be twenty-five millivolts, during the test, the A, B and C samples are placed in random order on the conveying system, the system is continuously operated and the decision and execution state is recorded, showing clear differentiated response phenomenon, when the A group sample passes, the system stably executes the regular flushing of the low-pressure cleaning mechanism because it does not detect the near-infrared signal below the organic matter light absorption threshold, when the B group sample passes, the system consistently determines the strong flushing condition and executes the standard single-pulse high-pressure flushing action, and when the C group sample passes, the system also starts the high-pressure cleaning mechanism first, but at the initial stage of the contact of high-pressure water flow with the green onion body, the data acquisition system generally records that the alternating fluctuation component instantaneously exceeds the vibration amplitude threshold value of twenty-five millivolts, and it is observed that the control system almost at the same time switches the cleaning mode to the gradient cleaning sequence, for specific test data, see Table 1.

[0043] Table 1: Differentiated cleaning and adaptive energy regulation test data table.

[0044]

[0045] Analysis of the data shown in Table 1 shows that when samples B-01 and C-01 faced exactly the same severe contamination, the system correctly initiated high-pressure cleaning. However, it was precisely because the latter produced an AC fluctuation component far exceeding the threshold under impact, a measurable physical phenomenon, that triggered the adaptive energy regulation mechanism based on mechanical response, causing the final execution mode to switch from a single-pulse high-voltage to a gradient cleaning sequence. As a direct result of this mode switch, the cell damage rate of Group C samples was successfully controlled to the same order of magnitude as that of Group B samples, far lower than that of the control group without intervention by this mechanism. At the same time, its post-cleaning residual rate was maintained at less than 1%, which is direct evidence of the effectiveness of this inherent mechanism.

[0046] Example 3: This example combines Figures 1 to 4 , a green onion processing and cleaning method integrating visual recognition and remote monitoring is described. Figure 1 As shown, Figure 1 Typical attenuation curves of the near-infrared reflection signal voltage (mV) versus time (ms) for healthy green onions (Group H) and damaged green onions (Group D) during the natural evaporation of the water film on their surfaces after cleaning are shown. These curves reflect the differences in near-infrared reflection characteristics between the two groups of samples at different time points. The solid line indicates the healthy green onions (Group H), whose near-infrared reflection signal voltage decreases rapidly over time, representing rapid evaporation of the water film. The dashed line indicates the damaged green onions (Group D), whose reflection signal decreases more slowly, reflecting increased hydrophilicity and a thicker residual water film due to tissue damage. The optimal detection time is also clearly marked by a vertical gray dashed line. This time point is determined by analyzing the maximum distinguishability of the near-infrared reflection signal voltages between the two groups of samples, i.e., the difference in maximum signal separation. This time point serves as the judgment window for determining the presence of quality defects, thus providing a highly reliable time point for subsequent automatic rejection using near-infrared optical properties. This graph, a key component of the quality screening mechanism derived from physical traces, embodies the core technical approach of the present invention, which accurately identifies hidden agricultural product damage through signal attenuation without the need for additional hardware.

[0047] like Figure 2As shown, the process is first completed by the cleaning system to start the delay timer after the cleaning action, enter the waiting 30-80ms set delay stage to ensure the difference of water film evaporation, after the difference of water film evaporation is formed, the secondary detection is triggered, the second photosensitive sensing unit reacquires the reflected signal voltage, if it is healthy green onion detection, the output reflected intensity signal is thin water film reflected signal (high), when the signal is higher than the signal water film residual threshold value, the fourth voltage comparator judges, outputs no rejection instruction, the rejection device does not act, through quality inspection, the green onion is confirmed as healthy green onion and is moved to the healthy green onion path; if it is defective green onion detection, the output reflected intensity is thick water film reflected signal (low), the signal is lower than the signal water film residual threshold value, the fourth voltage comparator outputs sends the rejection instruction, the rejection device immediately executes the rejection action, the green onion is confirmed as strong hydrophilicity, thick water film, which indicates that it has signs of tissue damage or early corruption, and is moved out of the main line to the defective green onion path.

[0048] As shown in the flowchart, Figure 3 , Figure 3 The comparative curve of water pressure (unit: MPa) changing with time (unit: ms) in the two cleaning modes of single pulse high pressure cleaning and gradient cleaning sequence in the application is shown, which embodies the difference characteristics of the cleaning energy control strategy. The solid line shown is single pulse high pressure cleaning, which is 0 MPa at 0 ms initially, rises to about 1.45 MPa at 100 ms, and maintains constant high pressure between 100 ms and 250 ms, and quickly drops back to 0 MPa after 250 ms. The whole is a single, high-intensity, fixed-duration pulse water flow. The dashed line shown is the gradient cleaning sequence, which realizes the gradual release of water pressure through segmented control. Specifically, it rises to about 0.75 MPa at about 50 ms as a pre-wetting pulse, and drops to zero after 150 ms, and then rises to about 1.45 MPa as a main impact segment, and drops to zero after 250 ms. This sequence effectively avoids structural damage to the fragile onion body through pressure staging.

[0049] As shown in the flowchart, Figure 4 , the process starts at the start node, first executes the high pressure cleaning mechanism start, then real-time acquires the second photosensitive sensing unit signal, the system extracts the alternating current fluctuation component through the high pass filter to obtain the signal waveform reflecting the stress state of the green onion body. The process corresponds to the vibration response waveform shown in the figure, and the characteristic quantity is the alternating current component (alternating current component). The system compares the with a vibration threshold value determined by the system through a pre-vibration threshold value, if the judgment result is If the value of the amplitude of the AC component of the vibration signal is greater than the vibration threshold value, it indicates that the current impact has a risk of damage, and the system switches to the gradient cleaning sequence; the gradient cleaning sequence consists of two pulses, where pulse 1: 50% pressure, pulse 2: 100% pressure, and the corresponding times are If the value of the amplitude of the AC component of the vibration signal is less than the vibration threshold value, it indicates that the current impact has no risk of damage, and the system switches to the single-pulse high-pressure cleaning sequence; the single-pulse high-pressure cleaning sequence consists of one pulse, where the pressure is 100%, and the time is the standard time, and the cleaning operation is completed at the standard high pressure. If the value of the amplitude of the AC component of the vibration signal is greater than the vibration threshold value, it indicates that the current impact has a risk of damage, and the system switches to the gradient cleaning sequence; the gradient cleaning sequence consists of two pulses, where pulse 1: 50% pressure, pulse 2: 100% pressure, and the corresponding times are If the value of the amplitude of the AC component of the vibration signal is less than the vibration threshold value, it indicates that the current impact has no risk of damage, and the system switches to the single-pulse high-pressure cleaning sequence; the single-pulse high-pressure cleaning sequence consists of one pulse, where the pressure is 100%, and the time is the standard time, and the cleaning operation is completed at the standard high pressure. If the value of the amplitude of the AC component of the vibration signal is greater than the vibration threshold value, it indicates that the current impact has a risk of damage, and the system switches to the gradient cleaning sequence; the gradient cleaning sequence consists of two pulses, where pulse 1: 50% pressure, pulse 2: 100% pressure, and the corresponding times are

[0050] In a production scenario of processing a specific category of organic green onions with fragile skin texture, to further reduce the microscopic physical damage rate and improve the accuracy of subsequent quality defect removal based on water film residue, it is necessary to optimize the energy structure of the gradient cleaning sequence and the time delay and judgment threshold in quality detection, which are two key dynamic links in the system, more precisely and data-drivenly.

[0051] For the optimization of the gradient cleaning sequence, the goal is to establish an energy release model that accurately associates response strength with risk level. To this end, the following procedure is adopted: when high-pressure cleaning is triggered, and the system monitors the AC fluctuation component in real time When the amplitude of the AC component of the vibration signal exceeds the preset vibration amplitude threshold value , the system executes a double-pulse cleaning sequence. The first pre-wetting pulse of this sequence is set to 50% of the standard high pressure, and its duration is determined by the relationship , where is a reference time, and the second main impact pulse has a pressure of standard high pressure and a duration of . According to this relationship, as the vibration of the onion body intensifies, i.e. the numerical value increases, the duration of the pre-wetting pulse shortens accordingly, thereby adaptively adjusting the total energy input and action time. To avoid the setting uncertainty caused by the mutual dependence between the predetermined time delay and the water film residue detection threshold in the quality detection link, this embodiment introduces an offline joint calibration experiment procedure. The initial state of this procedure is defined as follows: prepare two groups of samples, H group for healthy green onions confirmed to have no damage, and D group for green onions with hidden bruises made on the surface through standardized micro-impact. The execution steps of the procedure are as follows: wet both groups of samples under the same conditions and immediately place them in the detection range of the second photosensitive sensing unit. From The near-infrared reflectance intensity signal voltage of each sample is continuously collected and recorded at five-millisecond intervals from the time point t0 until the signal reaches stability, thereby obtaining a voltage-time decay curve representing the evaporation process of the water film on the surface of each sample.

[0052] The analysis of the collected data is as follows: since the damaged tissue of the samples in group D has stronger hydrophilicity, the water film evaporates more slowly than that of group H, which is manifested by the fact that the decay of the voltage-time curve of the samples in group D is more gentle. To find the peak point that distinguishes the signal difference between the two groups of samples, a time-dependent signal separation function is defined as follows: , which is the difference between the average voltage of the samples in group H at time t1 and the average voltage of the samples in group D at time t2 , i.e. ; through calculation, the time point t3 at which the value of f(t) reaches the maximum value is found, which is determined as a predetermined time delay optimized by data. Once t3 is determined, the water film residual detection threshold value t4 can be set at the intermediate value of the voltages of the two groups of samples at this time point, i.e. ; through this joint calibration procedure, the setting process of the two interrelated parameters is converted into a data-driven optimization problem.

[0053] In order to ensure that the technical solution of the present application can maintain stable design performance in different regions, different seasons, and when facing different batches of raw materials, a standardized pre-calibration and baseline establishment procedure can be performed when the system is first deployed or when the processing object is replaced. This procedure first requires the operator to select a group of reference samples from the current batch to be processed, which must include healthy and clean scallions, scallions with typical stains of the batch, and scallions with the most fragile texture. After the system is operated under no load and the background reflectance signal of the conveying belt is recorded, the reference samples are passed through the optical detection module one by one to complete the signal collection of visible light and near-infrared band reflectance intensity in the non-cleaning state, respectively. Based on the data collected by the system, a set of initial working thresholds for the current batch is automatically calculated and set, in which the scallion presence threshold is set at the intermediate value of the visible light reflectance intensity signals of the clean scallions and the conveying belt background, and the organic matter light absorption threshold is established based on the near-infrared absorption characteristics of the typical stain sample. At the same time, by applying a standard test water flow pulse to the sample with the most fragile texture, the vibration response data can be obtained to provide a reference for setting the vibration amplitude threshold.

[0054] ​​​​​In order to further enhance the long-term self-adaptive ability of the system to environmental and raw material changes, and improve the recognition robustness of new or atypical organic stains, the system also integrates a dynamic updating and maintenance mechanism of the stain optical feature library. During the regular processing, the system will continuously record the 950 nanometer near-infrared reflection intensity signal values corresponding to all events triggering the high-pressure cleaning mechanism, and store them in the database. During the system maintenance period, a background analysis program will perform statistical cluster analysis on the signal values accumulated in the database. The analysis aims to identify new signal distribution clusters that have significant statistical differences with the data in the existing feature library. If such new clusters are detected, the system will prompt the operator to extract physical samples associated with the spatial and temporal position of the abnormal signal from the production line for inspection. Once confirmed as a new stain by manual, the near-infrared reflection feature data of the sample will be officially included in the core stain optical feature library of the system, and the organic matter absorption threshold will also be recalculated and adjusted based on the updated feature library, thereby realizing the long-term stability of the system's adaptability to environmental changes and its recognition accuracy. The remote monitoring function and dynamic maintenance procedure of the stain feature library disclosed in the present application are as follows. The remote monitoring of the system is realized through an industrial Ethernet module integrated into the main controller. The module uses the OPC-UA protocol to push data to the local server or cloud platform. The data presented in real-time by the remote terminal interface of the system includes the current batch of green onions processing count, the minute trigger frequency of the high and low pressure cleaning mechanism, the activation ratio of the gradient cleaning sequence, and the real-time rejection rate of defective products. At the same time, the interface authorizes users with secondary or higher permissions to remotely call the pre-set threshold parameter group after changing the processing batch. In addition, the system records all near-infrared reflection intensity signal values and their corresponding AC fluctuation component peak values that trigger the high-pressure cleaning event. When the background analysis program identifies new data clusters that have significant statistical differences with the existing stain optical feature library using clustering algorithms, the remote monitoring interface will alert the maintenance personnel and highlight the timestamp and production line location information of the physical sample corresponding to the data cluster, in order to guide them to perform physical sampling and feature library updating.

[0055] It is apparent for those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and can be implemented in other specific forms without departing from the spirit or essential characteristics of the present application.

[0056] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application.

Claims

1. A green onion processing and cleaning method integrating visual recognition and remote monitoring, characterized in that: The following steps are involved: Step a, obtaining a reflection intensity signal of the green onion surface to the visible light band on the green onion conveying path; and simultaneously obtaining a reflection intensity signal of the green onion surface to the near-infrared band of the central wavelength; Step b: inputting the reflection intensity signal of the near-infrared band into a first voltage comparator, and when the reflection intensity signal of the near-infrared band is lower than the organic matter absorption threshold value determined by pre-calibration, the first voltage comparator outputs a first logic high level signal; Step c, inputting the reflection intensity signal of the visible light band into a second voltage comparator, and when the reflection intensity signal of the visible light band is lower than a threshold value of the presence of green onions determined by pre-calibration, the second voltage comparator outputs a second logic high level signal; Step d, based on the AND logic judgment of the first logic high level signal and the second logic high level signal, driving the high-pressure cleaning mechanism to perform a strong wash on the green onions, wherein the strong wash is performed when the first logic high level signal is true and the second logic high level signal is true; Step e: when the first logic high level signal is false and the second logic high level signal is true, driving the low-pressure cleaning mechanism to perform conventional washing on the green onions; The reflection intensity signal of the visible light band is obtained by the first photosensor unit; the reflection intensity signal of the near-infrared band with a central wavelength of 950 nanometers is obtained by the second photosensor unit, and the second photosensor unit is only sensitive to near-infrared light with a central wavelength of 950 nanometers.

2. A green onion processing and cleaning method integrating visual recognition and remote monitoring as claimed in claim 1, characterized in that: The high-pressure cleaning mechanism and the low-pressure cleaning mechanism are both independent water paths controlled by solenoid valves. The first logic high level signal and the second logic high level signal directly control the opening and closing of the solenoid valves through the logic gate circuit.

3. A green onion processing and cleaning method integrating visual recognition and remote monitoring as claimed in claim 2, characterized in that: The step of strong flushing also includes: obtaining the reflection intensity signal of the second photosensor unit at the same time as the high-pressure cleaning mechanism starts flushing; extracting the AC fluctuation component of the reflection intensity signal through a high-pass filter; inputting the AC fluctuation component into a third voltage comparator, and when the AC fluctuation component exceeds the vibration amplitude threshold, adjusting the energy release mode of the high-pressure cleaning mechanism based on the AC fluctuation component; the energy release mode is to adjust the single cleaning action of the high-pressure cleaning mechanism to a gradient cleaning sequence consisting of at least two pulses with increasing pressure or increasing duration.

4. A green onion processing and cleaning method integrating visual recognition and remote monitoring as claimed in claim 3, characterized in that: The adjustment condition of the energy release mode is that the difference between the AC fluctuation component and the AC fluctuation component of the reflection intensity signal of the first photosensor unit exceeds a differential fluctuation threshold value determined by pre-calibration.

5. A green onion processing and cleaning method integrating visual recognition and remote monitoring as claimed in claim 4, characterized in that: A vortex cavitation nozzle is provided at the end of the high-pressure cleaning mechanism. The vortex cavitation nozzle has an internal channel structure. The internal channel structure causes the local pressure of the high-pressure water flow passing through to drop below the saturated vapor pressure of water, thereby forming cavitation microjets on the surface of the green onion.

6. A green onion processing and cleaning method integrating visual recognition and remote monitoring as claimed in claim 5, characterized in that: The internal channel structure of the vortex cavitation nozzle includes a spirally tapered channel, which increases the water flow velocity, resulting in a local pressure drop, satisfying the Bernoulli principle, that is: ,in, is the local pressure of the water flow, is the local velocity of the water flow, is the water flow density, is the energy constant in Bernoulli's principle.

7. A green onion processing and cleaning method integrating visual recognition and remote monitoring as claimed in claim 6, characterized in that: The water pressure of the high-pressure cleaning mechanism is higher than 1.5 MPa.

8. A green onion processing and cleaning method integrating visual recognition and remote monitoring as claimed in claim 7, characterized in that: The following steps are also included: After a predetermined time delay after each graded cleaning action is completed, the reflection intensity signal of the near-infrared band with a central wavelength of 950 nanometers of the second photosensor unit is obtained again; and the again obtained reflection intensity signal is input into the fourth voltage comparator. When the signal is lower than the water film residual detection threshold determined by calibration of the epidermis of healthy green onions, the fourth voltage comparator outputs a rejection instruction, which is used to indicate that the green onions have quality defects and drive the rejection device to remove the green onions from the production line.

9. A green onion processing and cleaning method integrating visual recognition and remote monitoring as claimed in claim 8, characterized in that: The predetermined time delay ranges from thirty milliseconds to eighty milliseconds.

Images