Gold electrolytic recovery method and device based on AI vision

Abstract

The application belongs to the technical field of metal recovery, and provides a gold electrolytic recovery method and device based on AI vision, which can transfer the pretreated gold ion-containing electrolytic mother liquor to an electrolytic cell provided with a deposition plate, deposit gold on the deposition area on the surface of the deposition plate, continuously collect images of the deposition area by a camera unit, calculate the dense gold layer coverage rate and deposition defect area ratio of the deposition area by using a detection model, realize dynamic adjustment of the electrolysis parameters, and stop electrolysis and complete the current deposition when the preset completion condition is met. The scheme realizes non-contact real-time quantitative detection of the state of the gold deposition layer during the electrolysis process, can dynamically adjust the electrolysis parameters based on the model detection coverage rate, and preferentially triggers the inhibition operation when the defect area exceeds the standard, ensures that the deposition process is carried out in an excellent state, improves the consistency of the deposition product, reduces the requirements for the experience and concentration of the operator, and reduces human errors.

Description

Technical Field

[0001] This invention belongs to the field of metal recycling technology, specifically relating to a gold electrolytic recycling method and apparatus based on AI vision. Background Technology

[0002] Gold, a rare and valuable precious metal, is widely used in industries such as electronics, jewelry, electroplating, and chemicals, resulting in a large amount of gold-containing waste. The efficient and high-purity recovery of gold from this waste has significant economic and environmental value. Electrolytic deposition is one of the mainstream processes for gold recovery. Its basic principle is to place a solution containing gold ions in an electrolytic cell, and then use an external current to reduce the gold ions to elemental gold on the cathode surface.

[0003] However, traditional electrolytic recovery processes have many limitations, such as reliance on manual experience and difficulty in precise online control. Operators usually rely on fixed process parameters and manual observation to judge whether the process is normal. When defects such as rough deposits, blackening, or dendrite formation are found, irreversible quality problems have often already occurred. At this point, it is too late to make adjustments, and rework is generally required. Moreover, the quality of the entire process is highly dependent on the experience and focus of the technicians. This not only involves high labor intensity but also easily introduces subjective errors, which is not conducive to the standardization and large-scale stable production of the process. Summary of the Invention

[0004] To overcome the shortcomings of the prior art, this invention proposes a gold electrolytic recovery method based on AI vision, the method comprising: Gold-containing waste is pretreated in a pretreatment tank to generate an electrolyte mother liquor rich in gold ions, and the volume of the electrolyte mother liquor and the initial gold ion concentration in the electrolyte mother liquor are measured. The mother liquor is transferred to an electrolytic cell equipped with a deposition plate. Electrolysis is started with preset electrolysis parameters and a protective gas is introduced. Gold is then deposited in the deposition area on the surface of the deposition plate, and images of the deposition area are continuously acquired by a camera unit. The acquired image is input into the pre-trained detection model. The detection model calculates the dense gold layer coverage and the deposition defect area ratio of the deposition area. The electrolysis parameters are dynamically adjusted based on the dense gold layer coverage. When the deposition defect area ratio exceeds a preset threshold, defect suppression operation is performed first until the defect area ratio is not higher than the preset threshold. The amount of deposited charge in the electrolytic cell is continuously acquired. When the coverage of the dense gold layer reaches the preset target coverage and is maintained for no less than the preset time, or when the cumulative amount of deposited charge reaches the first proportion of the theoretical charge value corresponding to the initial gold ion concentration in the electrolytic mother liquor, electrolysis is stopped and the power supply is cut off to complete this deposition. The first proportion is no greater than 1.

[0005] Specifically, the pretreatment tank is equipped with a vibrating screen and a gas distributor, and the pretreatment of the gold-containing waste in the pretreatment tank includes: The gold-containing electronic waste is physically crushed and sorted to obtain a number of gold-plated waste particles, and each of the gold-plated waste particles is put into the vibrating screen of the pretreatment tank; Mechanical vibration and electric heating are applied to the medium in the pretreatment tank, and gas is introduced into the medium through the gas distributor to enhance the disturbance and carry out the enhanced stripping reaction, so that gold ions can fully enter the solution to form the electrolytic mother liquor.

[0006] Specifically, the method for training the detection model includes: Multiple combinations of electrolysis parameters were constructed and electrolyzed with mother liquors of different gold ion concentrations. Multiple images of the deposition area in the deposition plate were acquired by a camera unit, and the electrolysis parameter combinations and gold ion concentrations corresponding to each image were recorded. Pixel-level semantic segmentation and annotation are performed on each of the images, and each pixel in the image is labeled as a dense gold layer, dendritic defect, void defect or substrate uncovered, forming an original labeled dataset; The original labeled dataset is divided into a training set, a validation set, and a test set. The images in the training set and the validation set are subjected to data augmentation processing, including standardization, random cropping, and flipping. The detection model is trained using the processed training set, and the performance of the detection model is monitored using the processed validation set. The average intersection-union ratio (IUR) of the model for segmenting the four types of regions is calculated using the test set to evaluate the performance of the detection model. When the average IUR is not lower than a preset value, the detection model is considered to have completed training.

[0007] Further, the electrolysis parameters include the electrolysis current density and the gas flow rate of the protective gas, and the dynamic adjustment of the electrolysis parameters based on the dense gold layer coverage includes: The dense gold layer coverage rate output by the detection model is obtained at a predetermined frequency, and the deviation and rate of change of the deviation between the real-time dense gold layer coverage rate and the expected value of the preset target coverage rate curve at the current moment are calculated. The deviation and rate of change of the real-time dense gold layer coverage from the target coverage are input into the fuzzy controller. The fuzzy controller performs inference and defuzzification calculations based on the preset fuzzy rule base and outputs the first adjustment amount corresponding to the electrolysis current density and the second adjustment amount corresponding to the gas flow rate. The electrolysis current density is updated based on the first adjustment amount, with a rate range of 0.5 A / m²·min to 5 A / m²·min as the limit, and the gas flow rate is updated based on the second adjustment amount, with a rate range of 0.3 L / min·L to 1.0 L / min·L as the limit.

[0008] Preferably, the step of preferentially performing a defect suppression operation when the deposition defect area ratio exceeds a preset threshold, until the defect area ratio is not higher than the preset threshold, includes: If the deposition defect area ratio is detected to exceed the preset threshold, the dynamic adjustment of the electrolysis parameters is stopped, and the electrolysis current density is reduced by 20%-50% accordingly, or a reverse pulse current with an amplitude of 10%-30% of the current electrolysis current and a duration of 0.1s-1s is applied until the defect area ratio is not higher than the preset threshold.

[0009] Furthermore, the method also includes: After completing this deposition, acquire and record the dense gold layer coverage and cumulative deposited charge at the time when the deposition was triggered. If the current deposition is triggered by a condition corresponding to the dense gold layer coverage, the deposition plate is removed from the electrolytic cell and replaced with a new deposition plate; if the current deposition is triggered by a condition corresponding to the cumulative deposited charge, the liquid in the electrolytic cell is removed and new electrolyte is added. Repeat the steps of starting electrolysis with preset electrolysis parameters and introducing protective gas.

[0010] Preferably, the method further includes: If the current deposition is triggered by the condition of the corresponding dense gold layer coverage, and the cumulative deposited charge reaches the second ratio of the theoretical charge value corresponding to the initial gold ion concentration in the electrolytic mother liquor, the liquid in the electrolytic cell is simultaneously removed and new electrolytic mother liquor is added. If the current deposition is triggered by the condition corresponding to the cumulative deposited charge amount, and the dense gold layer coverage reaches the third proportion value of the target coverage, the deposition plate is simultaneously removed from the electrolytic cell and replaced with a new deposition plate; the second proportion value and the third proportion value are both not greater than 1, and the second proportion value is not greater than the first proportion value.

[0011] Preferably, the method further includes: After the deposition plate is removed from the electrolytic cell, the weight and composition of the gold-bearing deposits on the deposition plate are estimated based on the total deposition charge recorded during the deposition process and the visual morphological characteristics of the deposition area output by the detection model.

[0012] This invention also proposes a gold electrolytic recovery device based on AI vision, the device comprising: The pretreatment module is used to pretreat gold-containing waste in a pretreatment tank to generate an electrolyte mother liquor rich in gold ions, and to measure the volume of the electrolyte mother liquor and the initial gold ion concentration in the electrolyte mother liquor. The deposition acquisition module is used to transfer the electrolytic mother liquor to an electrolytic cell equipped with a deposition plate, start electrolysis with preset electrolysis parameters and introduce protective gas, thereby depositing gold in the deposition area on the surface of the deposition plate, and continuously acquiring images of the deposition area through a camera unit; The control module is used to input the acquired image into the pre-trained detection model, calculate the dense gold layer coverage and the deposition defect area ratio of the deposition area through the detection model, dynamically adjust the electrolysis parameters based on the dense gold layer coverage, and perform defect suppression operation first when the deposition defect area ratio exceeds a preset threshold until the defect area ratio is not higher than the preset threshold. The determination module is used to continuously acquire the amount of deposited charge in the electrolytic cell. When the coverage of the dense gold layer reaches the preset target coverage and is maintained for no less than the preset time, or when the cumulative amount of deposited charge reaches the first proportion of the theoretical charge value corresponding to the initial gold ion concentration in the electrolytic mother liquor, the electrolysis is stopped and the power supply is cut off to complete the current deposition. The first proportion is no greater than 1.

[0013] The present invention also proposes a computer-readable storage medium storing executable instructions that, when executed by a processor, implement the AI ​​vision-based gold electrolytic recovery method described above.

[0014] The present invention has at least the following beneficial effects: The proposed solution, through a camera unit and a pre-trained detection model, enables continuous quantitative analysis of the density and number of defects in the deposited layer during electrolysis. This changes the traditional operation mode that relies on manual experience and offline sampling inspection, making the process highly transparent. The dense gold layer coverage rate output by the detection model is used as a feedback signal to dynamically adjust the electrolysis parameters, and defect suppression is performed in combination with the deposition defect area ratio. This allows for timely intervention before defects expand, avoiding irreversible defects or scrap, and significantly reducing the product non-conforming rate. The overall process greatly reduces the reliance on the operator's experience, focus, and subjective judgment, which is conducive to stable production and large-scale promotion. Furthermore, the proposed solution employs a pretreatment tank equipped with a vibrating screen and a gas distributor, combining mechanical vibration, electric heating, and gas disturbance to enhance stripping. This achieves uniform dispersion and dynamic contact of waste particles in the reaction medium, preventing agglomeration, increasing the reaction surface area, and making the stripping reaction more complete and faster. It can more thoroughly release gold ions from the waste, improving the overall recovery rate from the source. By acquiring images under different electrolysis parameters and gold ion concentrations, it ensures that the training data can cover various working conditions that may be encountered in actual production, giving the model broad adaptability and high generalization ability. Through fine annotation of four different regions, the area and distribution of different morphological regions can be accurately quantified, providing accurate pixel-level basis for calculating coverage and defect ratio. The use of average crossover ratio as an evaluation index ensures the high accuracy of the model. Based on this, this solution employs a fuzzy controller to achieve intelligent and flexible process control, while limiting the adjustment rate to prevent drastic process fluctuations or loss of control caused by excessive single adjustment amplitude, ensuring the stability and safety of control. It also provides clear and efficient defect correction methods, ensuring the rapid elimination of existing micro-defects. Furthermore, the solution allows for selective replacement of deposition plates and electrolytic mother liquor according to actual conditions, reducing equipment downtime and improving overall equipment utilization and continuous production capacity. Moreover, it provides a rapid and non-destructive online method for estimating the weight and composition of gold deposits based on deposition charge and visual morphological characteristics, offering valuable reference data for production statistics, cost accounting, and subsequent refining processes.

[0015] Therefore, this invention proposes a gold electrolytic recovery method and device based on AI vision. The proposed solution realizes non-contact real-time quantitative detection of the state of the gold deposition layer during the electrolysis process. It can dynamically adjust the electrolysis parameters based on the coverage detected by the model, and preferentially trigger the suppression operation when the defect area exceeds the standard, so as to ensure that the deposition process is carried out in a good state, improve the consistency of the deposition products, reduce the requirements for the experience and concentration of the operators, and reduce human error. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 The flowchart of the gold electrolytic recovery method based on AI vision provided in Example 1; Figure 2 This is a flowchart of a method for pre-treating gold-containing waste. Figure 3 Flowchart of the method for training the detection model; Figure 4 A flowchart illustrating the method for dynamically adjusting electrolysis parameters; Figure 5 This is a module structure diagram of the gold electrolytic recovery device based on AI vision provided in Example 2. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0019] Various embodiments of the invention will be described more fully below. The invention may have various embodiments, and adjustments and changes may be made therein. However, it should be understood that there is no intention to limit the various embodiments of the invention to the specific embodiments disclosed herein, but rather the invention should be understood to cover all modifications, equivalents, and / or alternatives falling within the spirit and scope of the various embodiments of the invention.

[0020] In the following, the terms “comprising” or “may include” as used in various embodiments of the invention indicate the presence of the disclosed functions, operations, or elements, and do not limit the addition of one or more functions, operations, or elements. Furthermore, as used in various embodiments of the invention, the terms “comprising,” “having,” and their cognates are intended only to indicate a specific feature, number, step, operation, element, component, or combination of the foregoing, and should not be construed as primarily excluding the presence of one or more other features, numbers, steps, operations, elements, components, or combinations of the foregoing, or the possibility of adding one or more combinations of the foregoing.

[0021] In various embodiments of the invention, the expression "or" or "at least one of A and / or B" includes any combination or all combinations of the words listed simultaneously. For example, the expression "A or B" or "at least one of A and / or B" may include A, may include B, or may include both A and B.

[0022] The expressions used in the various embodiments of the present invention (such as "first," "second," etc.) may modify various constituent elements in the various embodiments, but do not limit the corresponding constituent elements. For example, the above expressions do not limit the order and / or importance of the elements. The above expressions are only used for the purpose of distinguishing one element from other elements. For example, a first user device and a second user device refer to different user devices, although both are user devices. For example, a first element may be referred to as a second element without departing from the scope of the various embodiments of the present invention, and similarly, a second element may also be referred to as a first element.

[0023] It should be noted that, in this invention, unless otherwise explicitly specified and defined, terms such as "installation," "connection," and "fixation" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0024] In this invention, those skilled in the art should understand that the terms indicating orientation or positional relationship in the text are based on the orientation or positional relationship shown in the accompanying drawings, and are only for the purpose of facilitating the description of this invention and simplifying the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this invention.

[0025] The terminology used in the various embodiments of the invention is for the purpose of describing particular embodiments only and is not intended to limit the various embodiments of the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of the invention pertain. The terms (such as those defined in a generally used dictionary) are to be interpreted as having the same meaning as in the context of the relevant technical field and are not to be interpreted as having an idealized or overly formal meaning, unless clearly defined in the various embodiments of the invention.

[0026] Example 1 Please see Figures 1-4 This embodiment proposes an AI vision-based gold electrolytic recovery method. This method effectively improves the electrolytic efficiency of gold, ensures improved gold deposition uniformity, and thus enhances the gold recovery rate and purity. Simultaneously, it enables full-process visualization and unmanned operation. The method specifically includes: S100: Pre-treat gold-containing waste in a pre-treatment tank to generate an electrolyte mother liquor rich in gold ions, and measure the volume of the electrolyte mother liquor and the initial gold ion concentration in the electrolyte mother liquor.

[0027] In this embodiment, step S100 can use online photometric analysis, electrochemical sensors, or inductively coupled plasma atomic emission spectrometry after sampling to determine the initial mass concentration of gold ions in the mother electrolyte. Simultaneously, the total theoretical charge required to reduce all gold ions in the mother electrolyte can be calculated as the theoretical charge value corresponding to the initial gold ion concentration in the mother electrolyte.

[0028] S200: The mother liquor is transferred to an electrolytic cell equipped with a deposition plate. Electrolysis is started with preset electrolysis parameters and protective gas is introduced. Gold is then deposited in the deposition area on the surface of the deposition plate, and images of the deposition area are continuously collected by a camera unit.

[0029] In this embodiment, the camera unit includes a high-resolution industrial camera. The method proposed in this embodiment enables the industrial camera to face the deposition area on the surface of the deposition plate through the viewing window and continuously acquire images of the deposition area in frames.

[0030] S300: Input the acquired image into the pre-trained detection model, calculate the dense gold layer coverage and the ratio of deposition defect area in the deposition area through the detection model, dynamically adjust the electrolysis parameters based on the dense gold layer coverage, and perform defect suppression operation first when the ratio of deposition defect area exceeds the preset threshold until the defect area ratio is not higher than the preset threshold.

[0031] In this embodiment, the deposition defect area ratio is equal to the ratio of the sum of the dendrite area and the pore area to the sum of the dense gold layer area, the dendrite area, and the pore area.

[0032] In this embodiment, the defect suppression operation specifically includes stopping the dynamic adjustment of electrolysis parameters and reducing the electrolysis current density by 20%-50%, or applying a reverse pulse current with an amplitude of 10%-30% of the current electrolysis current and a duration of 0.1s-1s, until the defect area ratio is not higher than a preset threshold.

[0033] S400: Continuously acquire the amount of deposited charge in the electrolytic cell. When the dense gold layer coverage reaches the preset target coverage and is maintained for no less than the preset time, or when the cumulative amount of deposited charge reaches the first proportion of the theoretical charge value corresponding to the initial gold ion concentration in the mother liquor, stop electrolysis and cut off the power supply to complete this deposition.

[0034] It should be noted that the preset time is 5 minutes, and the first ratio value is no greater than 1. In this embodiment, the first ratio value can be set to 90%.

[0035] Optionally, the method proposed in this embodiment can provide controllable DC power to the electrolytic cell using a high-precision DC power supply, connect a Hall sensor in series in the total circuit to measure the current value in real time, detect the voltage value in the electrolytic cell using a high-precision voltage sensor, and collect the signals from the Hall sensor and the voltage sensor at a fixed frequency through the analog input unit of the PLC, thereby realizing the calculation of the amount of deposited charge. For example, the method proposed in this embodiment can read the instantaneous current value transmitted by the Hall sensor based on a fixed sampling interval of 0.1 seconds, and continuously perform cumulative calculation using the rectangular method. That is, the current value at each sampling moment is multiplied by the sampling time interval to obtain the amount of tiny charge passing through in that short period of time. Then, all the tiny charge amounts are accumulated from the beginning to the present to obtain the value of the cumulative deposited charge.

[0036] Specifically, the pretreatment tank is equipped with a vibrating screen and a gas distributor. Step S100, which involves pretreating the gold-containing waste in the pretreatment tank, specifically includes: S110: Physically crush and sort the gold-containing electronic waste to obtain a number of gold-plated waste particles, and put each gold-plated waste particle into the vibrating screen of the pretreatment tank.

[0037] In one alternative embodiment, the vibrating screen comprises a screen made of titanium alloy with a mesh size of 2mm-10mm.

[0038] S120: Mechanical vibration and electric heating are applied to the medium in the pretreatment tank, and gas is introduced into the medium through a gas distributor to enhance the disturbance and carry out the enhanced stripping reaction, so that gold ions can fully enter the solution to form the electrolytic mother liquor.

[0039] Preferably, step S120 may further involve filtering and adjusting the mother electrolyte, and measuring the volume and initial gold ion concentration accordingly.

[0040] In this embodiment, the mechanical vibration adopts an asymmetric elliptical vibration trajectory. Step S120 applies mechanical vibration and electric heating, including mechanical vibration with a frequency of 5Hz-50Hz and electric heating with a heating temperature of 40℃-80℃. The time for strengthening the stripping reaction is 10min-60min. Preferably, the vibration frequency and electric heating temperature can be interlocked. The vibration frequency increases by 5%-10% for every 10℃ increase in temperature, and the vibration frequency decreases accordingly when the temperature decreases.

[0041] Specifically, methods for training the detection model include: S310: Construct various combinations of electrolysis parameters and electrolyze mother liquors with different gold ion concentrations. Acquire multiple images of the deposition area in the deposition plate through a camera unit and record the electrolysis parameter combinations and gold ion concentrations corresponding to each image.

[0042] Specifically, after recording is completed, step S310 will generate an image-parameter correspondence table containing images, electrolysis parameter combinations, and gold ion concentration values.

[0043] S320: Perform pixel-level semantic segmentation and annotation on each image, labeling each pixel in the image as a dense gold layer, dendritic defect, void defect, or uncovered substrate, forming the original labeled dataset.

[0044] S330: Divide the original labeled dataset into independent training, validation, and test sets. Perform data augmentation processing such as standardization, random cropping, and flipping on the images in the training and validation sets. Use the processed training set to train the detection model and monitor the performance of the detection model through the processed validation set.

[0045] Specifically, the detection model is built based on a convolutional neural network with an encoder-decoder structure. In step S330, the detection model is trained in a supervised manner using a processed training set with the goal of minimizing the pixel-level cross-entropy loss function. During the training process, a processed validation set is used to monitor the model's loss value and segmentation accuracy to prevent overfitting.

[0046] S340: The average crossover ratio (CRO) of the model for segmenting the four types of regions is calculated using the test set to evaluate the performance of the detection model. When the average CRO is not lower than the preset value, the detection model is deemed to have completed training.

[0047] It should be noted that the mean intersection over union (mIoU) is the most core and commonly used quantitative evaluation metric in computer vision. The mean IoU can objectively and accurately measure the ability of the detection model to identify different depositional morphologies. Step S340 calculates a separate IoU for the four types of regions: dense gold layer, dendritic defects, porosity defects, and uncovered substrate. The IoU of all the IoUs are then summed and averaged to obtain the mean IoU, thereby ensuring the accuracy of the detection model's prediction for each type of region.

[0048] In an optional implementation, the preset value can be set to 0.8 so that the detection model is considered to have completed training when the average crossover ratio is not less than 0.8.

[0049] Furthermore, the electrolysis parameters include the electrolysis current density and the gas flow rate of the protective gas. The dynamic adjustment of the electrolysis parameters based on the dense gold layer coverage described in step S300 includes: S350: Acquire the dense gold layer coverage output by the detection model at a predetermined frequency, and calculate the deviation and rate of change of the real-time dense gold layer coverage and the expected value of the preset target coverage curve at the current moment.

[0050] S360: Input the deviation and deviation change rate between the real-time dense gold layer coverage and the target coverage into the fuzzy controller. The fuzzy controller performs inference and defuzzification calculations based on the preset fuzzy rule base, and outputs the first adjustment amount corresponding to the electrolysis current density and the second adjustment amount corresponding to the gas flow rate.

[0051] When the current dense gold layer coverage is lower than the target value, the fuzzy controller will output a larger adjustment amount to increase the deposition rate and allow the gold layer to spread quickly in the deposition area; when the current dense gold layer coverage is close to the target value, the fuzzy controller will output a smaller adjustment amount to maintain a stable and gentle environment, ensuring that the deposition area not covered by the gold layer is filled in the densest way.

[0052] S370: The electrolysis current density is updated based on the first adjustment amount, with a rate range of 0.5A / m²·min-5A / m²·min as the limit, and the gas flow rate is updated based on the second adjustment amount, with a rate range of 0.3L / min·L-1.0L / min·L as the limit.

[0053] Furthermore, the method proposed in this embodiment also includes: S500: After completing this deposition, acquire and record the dense gold layer coverage and cumulative deposited charge at the time the deposition was triggered.

[0054] S600: If the current deposition is triggered by the condition corresponding to the dense gold layer coverage, the deposition plate is removed from the electrolytic cell and replaced with a new deposition plate; if the current deposition is triggered by the condition corresponding to the cumulative deposited charge, the liquid in the electrolytic cell is removed and new electrolyte is added.

[0055] Based on this, step S600 can also be preferably set as follows: if the current deposition is triggered by the condition of the corresponding dense gold layer coverage, and the cumulative deposited charge reaches the second ratio of the theoretical charge value corresponding to the initial gold ion concentration in the electrolytic mother liquor, the liquid in the electrolytic cell is simultaneously removed and new electrolytic mother liquor is added. If the deposition is triggered by the condition of the corresponding cumulative deposited charge and the dense gold layer coverage reaches the third proportion of the target coverage, the deposition plate is simultaneously removed from the electrolytic cell and replaced with a new deposition plate; the second proportion and the third proportion are both no greater than 1 and the second proportion is no greater than the first proportion.

[0056] After the deposition plate and / or liquid are replaced, the method proposed in this embodiment can be performed again in step S100 to start electrolysis with preset electrolysis parameters and introduce protective gas.

[0057] In step S600, when the deposition area of ​​the deposition plate is completed and there are still some gold ions in the electrolytic mother liquor, the completed deposition plate can be taken out and a new deposition plate can be used to continue electrolytic deposition. When the gold ions in the electrolytic mother liquor are exhausted or nearly exhausted, and the dense gold layer coverage in the deposition area of ​​the deposition plate is low, new electrolytic mother liquor can be added and the steps in step S200 of starting electrolysis with preset electrolysis parameters and introducing protective gas can be executed again to carry out a new round of deposition, so as to achieve continuous and automated batch production.

[0058] S700: After removing the deposition plate from the electrolytic cell, the weight and composition of the gold-bearing deposits on the deposition plate are estimated based on the total deposition charge recorded during the deposition process and the visual morphological characteristics of the deposition area output by the detection model.

[0059] It should be noted that the mass of the substance reacting on the electrode is proportional to the total amount of deposited charge in the electrolytic cell. Step S700 can calculate the theoretical mass of the deposited gold based on the total amount of deposited charge.

[0060] Preferably, the method proposed in this embodiment can also provide casting guidance to users based on the weight and composition of the gold-containing deposits on the deposition plate estimated in step S700.

[0061] For example, after the estimation is completed, a sedimentation smelting database is established, which includes historical sedimentation characteristics and their corresponding smelting parameter combinations and recovery rates. The weight and composition of the estimated gold-bearing sediments are used as sedimentation characteristics. Based on the sedimentation characteristics, historical data in the sedimentation smelting database are matched, and a recommended smelting parameter combination is obtained based on the matching results. The parameters in the smelting parameter combination may include, but are not limited to, smelting temperature parameters, flux type, flux ratio, smelting time and mold preheating temperature.

[0062] In an optional implementation, the method proposed in this embodiment can provide casting guidance through a trained smelting model. The smelting model is trained based on historical data in a sediment smelting database and can perform similarity matching or multi-objective optimization calculations between sediment features and data in the historical database. This provides users with at least two smelting strategies, one with the primary goal of improving the final gold ingot recovery rate and the other with the primary goal of improving the final gold ingot purity. Each smelting strategy corresponds to a different combination of smelting parameters, which can significantly reduce the dependence on operator experience and improve the overall recovery rate and product quality stability of gold from sediment to high-purity gold ingots.

[0063] Example 2 Please see Figure 5 This embodiment proposes an AI vision-based gold electrolytic recovery device to implement the AI ​​vision-based gold electrolytic recovery method proposed in Embodiment 1. The device includes: The pretreatment module 10 is used to pretreat gold-containing waste in a pretreatment tank to generate an electrolyte mother liquor rich in gold ions, and to measure the volume of the electrolyte mother liquor and the initial gold ion concentration in the electrolyte mother liquor. The deposition acquisition module 20 is used to transfer the electrolytic mother liquor to an electrolytic cell equipped with a deposition plate, start electrolysis with preset electrolysis parameters and introduce protective gas, thereby depositing gold in the deposition area on the surface of the deposition plate, and continuously acquiring images of the deposition area through the camera unit. The control module 30 is used to input the acquired image into the pre-trained detection model, calculate the dense gold layer coverage and the deposition defect area ratio of the deposition area through the detection model, dynamically adjust the electrolysis parameters based on the dense gold layer coverage, and perform defect suppression operation first when the deposition defect area ratio exceeds the preset threshold until the defect area ratio is not higher than the preset threshold. The determination module 40 is used to continuously acquire the amount of deposited charge in the electrolytic cell. When the dense gold layer coverage reaches the preset target coverage and is maintained for no less than the preset time, or when the cumulative amount of deposited charge reaches the first proportion of the theoretical charge value corresponding to the initial gold ion concentration in the mother liquor, the electrolysis is stopped and the power supply is cut off to complete the current deposition. The first proportion is no greater than 1.

[0064] Specifically, the pretreatment tank is equipped with a vibrating screen and a gas distributor. The pretreatment steps of the pretreatment module 10 include physically crushing and sorting the gold-containing electronic waste to obtain several gold-plated waste particles, and putting each gold-plated waste particle into the vibrating screen of the pretreatment tank. Mechanical vibration and electric heating are applied to the medium in the pretreatment tank, and gas is introduced into the medium through the gas distributor to enhance the disturbance and carry out the enhanced stripping reaction, so that gold ions can fully enter the solution to form an electrolyte mother liquor. The electrolyte mother liquor can also be filtered and adjusted, and the volume and initial gold ion concentration can be measured accordingly.

[0065] The control module 30 can acquire the dense gold layer coverage output by the detection model based on a predetermined frequency, calculate the deviation and rate of change of the real-time dense gold layer coverage and the preset target coverage curve at the current moment, and input the deviation and rate of change of the real-time dense gold layer coverage and the target coverage into the fuzzy controller. The fuzzy controller performs inference and defuzzification calculations according to the preset fuzzy rule base, and outputs a first adjustment amount corresponding to the electrolysis current density and a second adjustment amount corresponding to the gas flow rate. Then, with a rate range of 0.5A / m²·min-5A / m²·min as the limit, the electrolysis current density is updated based on the first adjustment amount, and with a rate range of 0.3L / min·L-1.0L / min·L as the limit, the gas flow rate is updated based on the second adjustment amount.

[0066] Furthermore, the apparatus proposed in this embodiment also includes: The acquisition module 50 is used to acquire and record the dense gold layer coverage and cumulative deposited charge at the time when the current deposition was triggered after the current deposition was completed. The circulation module 60 is used to remove the deposition plate from the electrolytic cell and replace it with a new deposition plate when the current deposition is triggered by the condition corresponding to the dense gold layer coverage; and to remove the liquid from the electrolytic cell and add new electrolyte when the current deposition is triggered by the condition corresponding to the cumulative deposited charge amount. The estimation module 70 is used to estimate the weight and composition of gold-bearing deposits on the deposition plate after the deposition plate is removed from the electrolytic cell, based on the total deposition charge recorded during the deposition process of the deposition plate in the electrolytic cell and the visual morphological characteristics of the deposition area output by the detection model.

[0067] Example 3 This embodiment also proposes a computer-readable storage medium storing computer instructions, which, when executed by a processor, implement the steps of the AI ​​vision-based gold electrolytic recovery method proposed in Embodiment 1 above.

[0068] It should be noted that computer-readable media, including both permanent and non-permanent, removable and non-removable media, can store information using any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, disk storage, quantum memory, graphene-based storage media or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0069] In summary, this invention proposes a gold electrolytic recovery method and apparatus based on AI vision. The proposed scheme achieves non-contact real-time quantitative detection of the state of the gold deposition layer during the electrolysis process. It can dynamically adjust the electrolysis parameters based on the coverage detected by the model, and preferentially trigger the suppression operation when the defect area exceeds the standard, ensuring that the deposition process is carried out in an excellent state, improving the consistency of the deposition products, reducing the requirements for the operator's experience and concentration, and reducing human error.

[0070] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An AI vision-based gold electrolytic recovery method, characterized in that, The method includes: Gold-containing waste is pretreated in a pretreatment tank to generate an electrolyte mother liquor rich in gold ions, and the volume of the electrolyte mother liquor and the initial gold ion concentration in the electrolyte mother liquor are measured. The mother liquor is transferred to an electrolytic cell equipped with a deposition plate. Electrolysis is started with preset electrolysis parameters and a protective gas is introduced. Gold is then deposited in the deposition area on the surface of the deposition plate, and images of the deposition area are continuously acquired by a camera unit. The acquired image is input into the pre-trained detection model. The detection model calculates the dense gold layer coverage and the deposition defect area ratio of the deposition area. The electrolysis parameters are dynamically adjusted based on the dense gold layer coverage. When the deposition defect area ratio exceeds a preset threshold, defect suppression operation is performed first until the defect area ratio is not higher than the preset threshold. The amount of deposited charge in the electrolytic cell is continuously acquired. When the coverage of the dense gold layer reaches the preset target coverage and is maintained for no less than the preset time, or when the cumulative amount of deposited charge reaches a first proportion of the theoretical charge value corresponding to the initial gold ion concentration in the mother liquor, electrolysis is stopped and the power supply is cut off to complete this deposition. The first proportion is no greater than 1. The methods for training the detection model include: Multiple combinations of electrolysis parameters were constructed and electrolyzed with mother liquors of different gold ion concentrations. Multiple images of the deposition area in the deposition plate were acquired by a camera unit, and the electrolysis parameter combinations and gold ion concentrations corresponding to each image were recorded. Pixel-level semantic segmentation and annotation are performed on each of the images, and each pixel in the image is labeled as a dense gold layer, dendritic defect, void defect or substrate uncovered, forming an original labeled dataset; The original labeled dataset is divided into a training set, a validation set, and a test set. The images in the training set and the validation set are subjected to data augmentation processing, including standardization, random cropping, and flipping. The detection model is trained using the processed training set, and the performance of the detection model is monitored using the processed validation set. The average intersection-union ratio (IUR) of the model for segmenting the four types of regions is calculated using the test set to evaluate the performance of the detection model. When the average IUR is not lower than a preset value, the detection model is considered to have completed training.

2. The AI vision-based gold electrolytic recovery method according to claim 1, characterized in that, The pretreatment tank is equipped with a vibrating screen and a gas distributor. The pretreatment of gold-containing waste in the pretreatment tank includes: The gold-containing electronic waste is physically crushed and sorted to obtain a number of gold-plated waste particles, and each of the gold-plated waste particles is put into the vibrating screen of the pretreatment tank. Mechanical vibration and electric heating are applied to the medium in the pretreatment tank, and gas is introduced into the medium through the gas distributor to enhance the disturbance and carry out the enhanced stripping reaction, so that gold ions can fully enter the solution to form the electrolytic mother liquor.

3. The gold electrolytic recovery method based on AI vision according to claim 1, characterized in that, The electrolysis parameters include the electrolysis current density and the gas flow rate of the protective gas. The dynamic adjustment of the electrolysis parameters based on the dense gold layer coverage includes: The dense gold layer coverage rate output by the detection model is obtained at a predetermined frequency, and the deviation and rate of change of the deviation between the real-time dense gold layer coverage rate and the expected value of the preset target coverage rate curve at the current moment are calculated. The deviation and rate of change of the real-time dense gold layer coverage from the target coverage are input into the fuzzy controller. The fuzzy controller performs inference and defuzzification calculations based on the preset fuzzy rule base and outputs the first adjustment amount corresponding to the electrolysis current density and the second adjustment amount corresponding to the gas flow rate. The electrolysis current density is updated based on the first adjustment amount, with a rate range of 0.5 A / m²·min to 5 A / m²·min as the limit, and the gas flow rate is updated based on the second adjustment amount, with a rate range of 0.3 L / min·L to 1.0 L / min·L as the limit.

4. The AI vision-based gold electrolytic recovery method according to claim 3, characterized in that, When the deposition defect area ratio exceeds a preset threshold, a defect suppression operation is preferentially performed until the defect area ratio is not higher than the preset threshold, including: If the deposition defect area ratio is detected to exceed the preset threshold, the dynamic adjustment of the electrolysis parameters is stopped, and the electrolysis current density is reduced by 20%-50% accordingly, or a reverse pulse current with an amplitude of 10%-30% of the current electrolysis current and a duration of 0.1s-1s is applied until the defect area ratio is not higher than the preset threshold.

5. The AI vision-based gold electrolytic recovery method according to claim 1, characterized in that, The method further includes: After completing this deposition, acquire and record the dense gold layer coverage and cumulative deposited charge at the time when the deposition was triggered. If the current deposition is triggered by a condition corresponding to the dense gold layer coverage, the deposition plate is removed from the electrolytic cell and replaced with a new deposition plate; if the current deposition is triggered by a condition corresponding to the cumulative deposited charge, the liquid in the electrolytic cell is removed and new electrolyte is added. Repeat the steps of starting electrolysis with preset electrolysis parameters and introducing protective gas.

6. The AI vision-based gold electrolytic recovery method according to claim 5, characterized in that, The method further includes: If the current deposition is triggered by the condition of the corresponding dense gold layer coverage, and the cumulative deposited charge reaches the second ratio of the theoretical charge value corresponding to the initial gold ion concentration in the electrolytic mother liquor, the liquid in the electrolytic cell is simultaneously removed and new electrolytic mother liquor is added. If the current deposition is triggered by the condition corresponding to the cumulative deposited charge amount, and the dense gold layer coverage reaches the third proportion value of the target coverage, the deposition plate is simultaneously removed from the electrolytic cell and replaced with a new deposition plate; the second proportion value and the third proportion value are both not greater than 1 and the second proportion value is not greater than the first proportion value.

7. The gold electrolytic recovery method based on AI vision according to claim 5 or 6, characterized in that, The method further includes: After the deposition plate is removed from the electrolytic cell, the weight and composition of the gold-bearing deposits on the deposition plate are estimated based on the total deposition charge recorded during the deposition process and the visual morphological characteristics of the deposition area output by the detection model.

8. An AI vision-based gold electrolytic recovery device, characterized in that, The apparatus for implementing the AI ​​vision-based gold electrolytic recovery method as described in any one of claims 1-7 includes: The pretreatment module is used to pretreat gold-containing waste in a pretreatment tank to generate an electrolyte mother liquor rich in gold ions, and to measure the volume of the electrolyte mother liquor and the initial gold ion concentration in the electrolyte mother liquor. The deposition acquisition module is used to transfer the electrolytic mother liquor to an electrolytic cell equipped with a deposition plate, start electrolysis with preset electrolysis parameters and introduce protective gas, thereby depositing gold in the deposition area on the surface of the deposition plate, and continuously acquiring images of the deposition area through a camera unit; The control module is used to input the acquired image into the pre-trained detection model, calculate the dense gold layer coverage and the deposition defect area ratio of the deposition area through the detection model, dynamically adjust the electrolysis parameters based on the dense gold layer coverage, and perform defect suppression operation first when the deposition defect area ratio exceeds a preset threshold until the defect area ratio is not higher than the preset threshold. The determination module is used to continuously acquire the amount of deposited charge in the electrolytic cell. When the coverage of the dense gold layer reaches the preset target coverage and is maintained for no less than the preset time, or when the cumulative amount of deposited charge reaches the first proportion of the theoretical charge value corresponding to the initial gold ion concentration in the electrolytic mother liquor, the electrolysis is stopped and the power supply is cut off to complete the current deposition. The first proportion is no greater than 1.

9. A computer-readable storage medium, characterized in that, It stores executable instructions for use by a processor to implement the AI ​​vision-based gold electrolytic recovery method as described in any one of claims 1-7.

Images