An intelligent medicine-taking recognition system and method based on machine vision

Abstract

This application relates to the technical field of health-related information systems and cloud platforms, and particularly relates to an intelligent medicine-taking recognition system and method based on machine vision, which collect medicine-taking data, the actual medicine-taking operation sequence of the medicine-taking robot, the actual medicine-taking cycle of the actual medicine-taking operation sequence, the historical medicine-taking operation sequence associated with the medicine-taking data, and the historical medicine-taking cycle of the historical medicine-taking operation sequence; filter out candidate medicine-taking operation sequences from the actual medicine-taking operation sequence; analyze whether there is an optimized sequence associated with the candidate medicine-taking operation sequence, and output an analysis log; when the analysis log indicates that there is an optimized sequence associated with the candidate medicine-taking operation sequence, the cloud platform generates medicine-taking sequences associated with multiple medicine-taking patients based on the optimized sequence and the candidate medicine-taking operation sequence; when the medicine-taking robot dispenses medicine to the medicine-taking patients, identity verification is performed. The present invention can plan the medicine-taking sequence closer and more time-saving compared with the actual sequence when planning the medicine-taking sequence in areas with complex patient conditions, improving the medicine-taking efficiency.

Description

Technical Field

[0001] This application relates to the technical field of health-related information systems and cloud platforms, and particularly to an intelligent medicine-taking recognition system and method based on machine vision. Background Art

[0002] With the development of computer technology, the demand for medicine-taking by patients is increasing.

[0003] In the current medicine-taking scheduling system, when a medicine-taking staff inputs multiple target patient wards and beds in the medicine-taking scheduling system, the medicine-taking scheduling system can generate the order to reach the corresponding target patient wards and beds. Compared with the past without a medicine-taking scheduling system, it not only facilitates the medicine-taking staff but also improves the medicine-taking efficiency.

[0004] However, in the existing medicine-taking scheduling system, when planning the medicine-taking order in areas with complex patient conditions, there will be problems that the planned order is farther and more time-consuming compared with the actual order, reducing the medicine-taking efficiency. Summary of the Invention

[0005] To achieve the above object, this application provides the following technical solutions:

[0006] According to the first aspect of the present invention, the present invention claims protection for an intelligent medicine-taking recognition method based on machine vision, which is applied to a cloud platform. The method includes:

[0007] The cloud platform collects medicine-taking data, and the medicine-taking data includes multiple medicine-taking patients;

[0008] The cloud platform collects the actual medicine-taking operation order of the medicine-taking robot, the actual medicine-taking cycle of the actual medicine-taking operation order, the historical medicine-taking operation order associated with the medicine-taking data, and the historical medicine-taking cycle of the historical medicine-taking operation order;

[0009] The cloud platform filters out candidate medicine-taking operation orders from the actual medicine-taking operation order, where the order deviation value of the candidate medicine-taking operation order from the historical medicine-taking operation order is not less than a preset deviation threshold, and the candidate medicine-taking cycle of the candidate medicine-taking operation order is less than the historical medicine-taking cycle;

[0010] The cloud platform analyzes whether there is an optimized order associated with the candidate medicine-taking operation order and outputs an analysis log;

[0011] When the analysis log indicates that there is an optimized order associated with the candidate medicine-taking operation order, the cloud platform generates a medicine-taking order associated with the multiple medicine-taking patients based on the optimized order and the candidate medicine-taking operation order;

[0012] When the medicine dispensing robot dispenses medicine to the patient, it performs identity verification.

[0013] Further, it analyzes whether there is an optimized order associated with the candidate medicine dispensing operation sequence and outputs an analysis log, including:

[0014] When there is a pre-set patient in the pre-set ward where the POI is located in the candidate medicine dispensing operation sequence, it outputs an analysis log with an optimized order associated with the candidate medicine dispensing operation sequence;

[0015] When there is no pre-set patient in the pre-set ward where the POI is located in the candidate medicine dispensing operation sequence, it outputs an analysis log without an optimized order associated with the candidate medicine dispensing operation sequence.

[0016] Further, it analyzes whether there is an optimized order associated with the candidate medicine dispensing operation sequence and outputs an analysis log, including:

[0017] During the process of the medicine dispensing robot dispensing medicine from the medicine dispensing box according to the candidate medicine dispensing operation sequence, it collects the first patient of the medicine dispensing robot and the second patient of the medicine dispensing box;

[0018] When the medicine dispensing similarity between the first patient and the second patient exceeds the pre-set medicine dispensing similarity threshold, it outputs an analysis log with an optimized order associated with the candidate medicine dispensing operation sequence;

[0019] When the medicine dispensing similarity between the first patient and the second patient does not exceed the pre-set medicine dispensing similarity threshold, it outputs an analysis log without an optimized order associated with the candidate medicine dispensing operation sequence.

[0020] Further, the method further includes:

[0021] When the analysis log indicates that there is no optimized order associated with the candidate medicine dispensing operation sequence, it takes the candidate medicine dispensing operation sequence as the medicine dispensing operation sequence associated with the multiple medicine dispensing patients.

[0022] Further, the method further includes:

[0023] Based on the medicine dispensing order, it pushes guiding information to the processor so that the processor can provide animation information associated with the optimized order to the medicine dispensing robot based on the guiding information.

[0024] Further, after outputting the analysis log with an optimized order associated with the candidate medicine dispensing operation sequence, the method further includes: transmitting verification information to the processor so that the processor can inform the medicine dispensing robot to verify the analysis log based on the verification information and feedback the verification log to the cloud platform;

[0025] When the parsed log indicates an optimized order associated with the candidate medicine dispensing operation sequence, the cloud platform generates the dispensing order associated with the multiple medicine-taking patients based on the optimized order and the candidate medicine dispensing operation sequence, including:

[0026] When both the parsed log and the verification log indicate an optimized order associated with the candidate medicine dispensing operation sequence, generate the dispensing order associated with the multiple medicine-taking patients based on the optimized order and the candidate medicine dispensing operation sequence.

[0027] Further, when the preset patient is a traditional Chinese medicine patient, it at least includes:

[0028] Patients with phlegm turbidity blocking the collaterals, patients with blood stasis in the heart vessels, patients with qi stagnation in the chest, or patients with cold congealing the heart vessels;

[0029] When the preset patient is a Western medicine patient, it at least includes: patients with rhinitis, patients with sinusitis, and patients with pharyngitis.

[0030] Further, when the medicine dispensing robot dispenses medicine to the medicine-taking patient and conducts identity verification, it further includes:

[0031] When the medicine dispensing robot stops in front of the medicine-taking patient, scan the facial information of the medicine-taking patient;

[0032] The medicine dispensing robot compares the facial information with the information in the patient database to obtain the drug prescription information of the corresponding patient;

[0033] The medicine-taking patient obtains the dispensed medicine by himself / herself from the medicine dispensing robot. When the medicine dispensing robot detects that the dispensed medicine obtained by the medicine-taking patient is inconsistent with the drug prescription information of the corresponding patient, send a warning to the cloud platform;

[0034] The medicine dispensing robot conducts recycling processing on the dispensed medicine.

[0035] According to the second aspect of the present invention, the present invention claims to protect an intelligent medicine dispensing recognition system based on machine vision, which is characterized in that it is used to execute the intelligent medicine dispensing recognition method based on machine vision, and the system includes:

[0036] An acquisition unit, used to acquire medicine dispensing data, where the medicine dispensing data includes multiple medicine-taking patients, acquire the actual medicine dispensing operation sequence of the medicine dispensing robot, the actual medicine dispensing cycle of the actual medicine dispensing operation sequence, the historical medicine dispensing operation sequence associated with the medicine dispensing data, and the historical medicine dispensing cycle of the historical medicine dispensing operation sequence;

[0037] A filtering unit for filtering out a candidate medicine-taking operation sequence from the actual medicine-taking operation sequence, where the sequence deviation value between the candidate medicine-taking operation sequence and the historical medicine-taking operation sequence is not less than a preset deviation threshold, and the candidate medicine-taking cycle of the candidate medicine-taking operation sequence is less than the historical medicine-taking cycle;

[0038] An analysis unit for analyzing whether there is an optimized sequence associated with the candidate medicine-taking operation sequence and outputting an analysis log;

[0039] An optimization unit for, when the analysis log indicates that there is an optimized sequence associated with the candidate medicine-taking operation sequence, generating a medicine-taking sequence associated with multiple medicine-taking patients based on the optimized sequence and the candidate medicine-taking operation sequence.

[0040] This application relates to the technical field of intelligent medicine-taking, and in particular to an intelligent medicine-taking recognition system and method based on machine vision, which collect medicine-taking data, the actual medicine-taking operation sequence of a medicine-taking robot, the actual medicine-taking cycle of the actual medicine-taking operation sequence, the historical medicine-taking operation sequence associated with the medicine-taking data, and the historical medicine-taking cycle of the historical medicine-taking operation sequence; filter out a candidate medicine-taking operation sequence from the actual medicine-taking operation sequence; analyze whether there is an optimized sequence associated with the candidate medicine-taking operation sequence and output an analysis log; when the analysis log indicates that there is an optimized sequence associated with the candidate medicine-taking operation sequence, the cloud platform generates a medicine-taking sequence associated with multiple medicine-taking patients based on the optimized sequence and the candidate medicine-taking operation sequence; when the medicine-taking robot takes medicine for the medicine-taking patient, identity verification is performed. The present invention can plan a closer and more time-saving medicine-taking sequence compared with the actual sequence when planning the medicine-taking sequence in an area with complex patient conditions, improving the medicine-taking efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a schematic diagram of an application scenario of an intelligent medicine-taking recognition method based on machine vision claimed in an embodiment of this application;

[0042] Figure 2 It is a schematic diagram of a second application scenario of an intelligent medicine-taking recognition method based on machine vision claimed in an embodiment of this application;

[0043] Figure 3 It is a flowchart of the operation of an intelligent medicine-taking recognition method based on machine vision claimed in an embodiment of this application;

[0044] Figure 4 It is a structural block diagram of an intelligent medicine-taking recognition system based on machine vision claimed in an embodiment of this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0045] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0046] The terms "first", "second", and "third" in the present application are only used for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first", "second", and "third" may explicitly or implicitly include at least one of such features. In the description of the present application, the meaning of "a plurality" is at least two, such as two, three, etc., unless otherwise specifically defined. All directional indications (such as up, down, left, right, front, back...) in the embodiments of the present application are only used to explain the relative relationship between components and the movement conditions, etc. in a specific posture (as shown in the accompanying drawings). When the specific posture changes, the directional indications will also change accordingly. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products, or devices.

[0047] Referring to "embodiment" herein means that a specific feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the present application. The phrase does not necessarily refer to the same embodiment every time it appears in the specification, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0048] As Figure 1 shown, the figure is a schematic diagram of an application scenario of an intelligent drug-taking recognition method based on machine vision provided by an embodiment of the present application. In this application scenario, the drug-taking patients include a first drug-taking patient 101, a second drug-taking patient 102, a third drug-taking patient 103, and a fourth drug-taking patient 104. Based on the above patients, the drug-taking scheduling system can plan the forward order, such as R1, R2, and R3 in the figure. When the drug-taking robot runs the drug-taking box to the first drug-taking patient 101 to complete the delivery, it can run the drug-taking box to the second drug-taking patient 102 (the drug-taking running order is as Figure 1 shown by R1 in the figure), and after completing the drug-taking, it runs the drug-taking box to the third drug-taking patient 103 (the drug-taking running order is as Figure 1As shown in R2, after the medicine collection is completed, the medicine collection operation moves the medicine collection box to the fourth patient 104 for medicine collection (the medicine collection operation sequence is as shown in Figure 1 R3).

[0049] From Figure 1 the known medicine collection operation sequence shown, this medicine collection operation sequence has a crossing situation. For example, after the medicine collection robot completes the medicine collection for the third patient 103 for medicine collection, due to the baffle 202 set in the center of the ward passage, the medicine collection box needs to go around to the end of the passage and then turn right to reach the fourth patient 104 for medicine collection. From the section of the journey from the detour to the right turn, the medicine collection robot has already advanced once during the medicine collection process for the second patient 102 for medicine collection. The planning of the medicine collection operation sequence is unreasonable.

[0050] In fact, after the medicine collection robot completes the medicine collection for the second patient 102 for medicine collection, it can take the green channel 201 to collect medicine for the third patient 103 for medicine collection, and then walk back to the medicine collection box through the green channel on foot, and then operate the medicine collection box to turn right at the end of the front passage to collect medicine for the fourth patient 104 for medicine collection.

[0051] It can be seen that the sequence planned by the current medicine collection scheduling system has the problem of long cycle, which reduces the medicine collection efficiency.

[0052] In view of this, the embodiment of the present application provides an intelligent medicine collection recognition method based on machine vision. This method is applied to the cloud platform. This method includes that the cloud platform collects medicine collection data, and the medicine collection data includes multiple patients for medicine collection. The cloud platform can also collect the actual medicine collection operation sequence of the medicine collection robot, the actual medicine collection cycle of the actual medicine collection operation sequence, the historical medicine collection operation sequence associated with the medicine collection data, and the historical medicine collection cycle of the historical medicine collection operation sequence. Then, the candidate medicine collection operation sequence is filtered out from the actual medicine collection operation sequence. The candidate medicine collection operation sequence has a sequence deviation value not less than the preset deviation value threshold compared with the historical medicine collection operation sequence, and the candidate medicine collection cycle of the candidate medicine collection operation sequence is less than the historical medicine collection cycle. Then, it is analyzed whether there is an optimized sequence associated with the candidate medicine collection operation sequence, and the analysis log is output. When the analysis log indicates that there is an optimized sequence associated with the candidate medicine collection operation sequence, the medicine collection sequence associated with multiple patients for medicine collection is generated based on the optimized sequence and the candidate medicine collection operation sequence. After the medicine collection sequence is output, the medicine collection robot can operate the medicine collection box to collect medicine based on the actual medicine collection operation sequence in the medicine collection sequence, and collect medicine in a non-carried medicine collection box manner based on the optimized sequence in the medicine collection sequence, so as to have a high medicine collection efficiency.

[0053] Such as Figure 2As shown in the figure, this figure is a schematic diagram of another application scenario of an intelligent medicine-taking recognition method based on machine vision provided by an embodiment of the present application. By using the intelligent medicine-taking recognition method based on machine vision in the embodiment of the present application to optimize the original medicine-taking operation sequence ( Figure 1 the medicine-taking operation sequences R1, R2, and R3 therein), the medicine-taking sequence shown in Figure 2 is output. This medicine-taking sequence includes medicine-taking operation sequences J1, J2, and optimization sequences F1 and F2. Among them, the optimization sequence can be a non-medicine-taking operation sequence.

[0054] Among them, the medicine-taking robot first takes medicine and operates the medicine-taking box to reach the second medicine-taking patient 102 from the first medicine-taking patient 101 according to the medicine-taking operation sequence J1. After delivering the goods to the second medicine-taking patient 102, it walks in the way of the optimization sequence F1 through the green channel 201 to reach the third medicine-taking patient 103. After delivering the goods to the third medicine-taking patient 103, it walks again in the way of the optimization sequence F2 through the green channel 201 back to the medicine-taking box. Then it takes medicine and operates the medicine-taking box to reach the fourth medicine-taking patient 104 according to the medicine-taking operation sequence J2 and delivers the goods to the fourth medicine-taking patient 104.

[0055] Thus, it can be seen that by walking from the second medicine-taking patient 102 to the third medicine-taking patient 103, there is no need to carry the medicine-taking box to the end of the channel, nor to operate the medicine-taking box to walk a section of the ward channel that has been walked before. For specific patients (such as having a green channel in the ward channel), taking medicine in a specific way (such as walking) can reduce the medicine-taking cycle and improve the medicine-taking efficiency.

[0056] In order to make the technical solution of the present application clearer and easier to understand, the intelligent medicine-taking recognition method based on machine vision provided by the embodiment of the present application will be introduced below with reference to the accompanying drawings. As Figure 3 shown in the figure, this figure is a flowchart of an intelligent medicine-taking recognition method based on machine vision provided by an embodiment of the present application. This method can be executed by the cloud platform or jointly executed by the cloud platform and the processor. Below, taking the joint execution of the cloud platform and the processor as an example, this method will be introduced. This method includes:

[0057] S301. The cloud platform collects medicine-taking data.

[0058] The medicine-taking data includes multiple medicine-taking patients. In some examples, the merchant can transmit an order request to the cloud platform. The order request can carry the geographical patients of the merchant and the goods required. After receiving the order request, the cloud platform can parse the order request and then output the medicine-taking data, that is, multiple medicine-taking patients who need to take medicine.

[0059] S302. The actual medicine-taking operation sequence of the medicine-taking robot on the cloud platform, the actual medicine-taking cycle of the actual medicine-taking operation sequence, the historical medicine-taking operation sequence associated with the medicine-taking data, and the historical medicine-taking cycle of the historical medicine-taking operation sequence.

[0060] The actual medicine-taking operation sequence refers to the sequence in which the medicine-taking robot advances with the medicine-taking box during the process of taking medicine for the above-mentioned multiple medicine-taking patients. The actual medicine-taking cycle refers to the cycle required for the medicine-taking robot to take medicine for multiple medicine-taking patients according to the actual medicine-taking operation sequence. The historical medicine-taking operation sequence refers to the sequence in which the historical medicine-taking robot advances with the medicine-taking box during the process of taking medicine for the above-mentioned multiple medicine-taking patients in the historical period. The historical medicine-taking cycle refers to the cycle required for the historical medicine-taking robot to take medicine for multiple medicine-taking patients according to the historical medicine-taking operation sequence. When there are multiple historical medicine-taking operation sequences, the sequence with the shortest historical medicine-taking cycle can be selected from the multiple historical medicine-taking operation sequences.

[0061] In some examples, the actual medicine-taking operation sequence, the actual medicine-taking cycle, the historical medicine-taking operation sequence, and the historical medicine-taking cycle can be stored in the cloud platform to facilitate the cloud platform to collect these data.

[0062] It should be noted that the present application does not specifically limit the execution order of S301 and S302. In some examples, S302 can also be executed first, followed by S301, or S301 and S302 can be executed simultaneously.

[0063] S303. The cloud platform filters out the candidate medicine-taking operation sequence from the actual medicine-taking operation sequence.

[0064] The sequence deviation value between the candidate medicine-taking operation sequence and the historical medicine-taking operation sequence is not less than the preset deviation value threshold, and the candidate medicine-taking cycle of the candidate medicine-taking operation sequence is less than the historical medicine-taking cycle.

[0065] Among them, the actual medicine-taking operation sequence can be the current medicine-taking operation sequence, or the current medicine-taking operation sequence and the previous medicine-taking operation sequence. That is to say, the actual medicine-taking operation sequence can be one sequence or multiple sequences. After the cloud platform collects the actual medicine-taking operation sequence and the historical medicine-taking operation sequence, it can filter out the candidate medicine-taking operation sequence from the actual medicine-taking operation sequence.

[0066] In some examples, the cloud platform can compare the actual medicine-taking operation sequence and the historical medicine-taking operation sequence in turn, and then find the sequence whose sequence deviation value between the actual medicine-taking operation sequence and the historical medicine-taking operation sequence is not less than the preset deviation value threshold, and the sequence whose actual medicine-taking cycle is less than the historical medicine-taking cycle, as the candidate medicine-taking operation sequence.

[0067] The cloud platform can extract features from the actual medicine collection operation sequence and the historical medicine collection operation sequence respectively, output the actual sequence vector and the historical sequence vector, then calculate the similarity between the actual sequence vector and the historical sequence vector, characterize the sequence deviation value through the similarity, and characterize the preset deviation value threshold through the preset similarity threshold. When the similarity is not less than the preset similarity threshold, it is considered that the sequence deviation value is not less than the preset deviation value threshold.

[0068] S304. The cloud platform analyzes whether there is an optimized sequence associated with the candidate medicine collection operation sequence.

[0069] The cloud platform analyzes whether there is an optimized sequence associated with the candidate medicine collection operation sequence and outputs the analysis log. When the determination log indicates that there is an optimized sequence associated with the candidate medicine collection operation sequence, S305 is executed; when the analysis log indicates that there is no optimized sequence associated with the candidate medicine collection operation sequence, S306 is executed.

[0070] The embodiments of the present application provide various ways to analyze whether there is an optimized sequence. The first one is introduced below.

[0071] The cloud platform can analyze whether there are preset patients in the preset ward where the POI is located in the candidate medicine collection operation sequence. When there are preset patients in the preset ward where the POI is located in the candidate medicine collection operation sequence, it outputs an analysis log indicating that there is an optimized sequence associated with the candidate medicine collection operation sequence. When there are no preset patients in the preset ward where the POI is located in the candidate medicine collection operation sequence, it outputs an analysis log indicating that there is no optimized sequence associated with the candidate medicine collection operation sequence.

[0072] Among them, the POI in the candidate medicine collection operation sequence can be the medicine collection patient, and the preset ward can be customized. For example, the preset ward can be within a preset number of meters centered on the POI. The preset patient may not be convenient for the medicine collection box to move forward but is conducive to walking forward. For example, the preset patient can be a patient with phlegm turbidity blocking, blood stasis blocking the heart vessels, qi stagnation in the chest, or cold congealing the heart vessels; when the preset patient is a Western medicine patient, it can at least include: rhinitis patients, sinusitis patients, and pharyngitis patients.

[0073] For example, as Figure 2 shown, the candidate medicine collection operation sequence can be sequence J1 and sequence J2. In the overall sequence composed of sequence J1 and sequence J2, the POI includes the first medicine collection patient 101, the second medicine collection patient 102, and the fourth medicine collection patient 104. The cloud platform can sequentially analyze whether there are patients with phlegm turbidity blocking, blood stasis blocking the heart vessels, qi stagnation in the chest, or cold congealing the heart vessels within 30 meters of the first medicine collection patient 101. Obviously, there are no such preset patients within 30 meters of the first medicine collection patient 101. Then it continues to analyze whether there are patients with phlegm turbidity blocking, blood stasis blocking the heart vessels, qi stagnation in the chest, or cold congealing the heart vessels within 30 meters of the second medicine collection patient 102. FromFigure 2 It can be seen that there is a patient with blood stasis in the heart within 30 meters of the second patient 102 picking up medicine. Therefore, the cloud platform believes that there is an optimized order associated with the running order of picking up medicine for this waiting selection, and the optimized orders are order F1 and order F2. Then the cloud platform continues to analyze the subsequent patients picking up medicine to see if the preset ward where the patient picking up medicine is located includes preset patients, which will not be elaborated here.

[0074] It should be noted that the specific values in the above examples are only for exemplary introduction, and in other examples, they can also be other values.

[0075] The following introduces the second method of analyzing whether there is an optimized order associated with the running order of picking up medicine for waiting selection.

[0076] The cloud platform can collect the first patient of the medicine-picking robot and the second patient of the medicine box during the process of the medicine-picking robot picking up medicine from the medicine box in accordance with the running order of picking up medicine for waiting selection. Then calculate the similarity of picking up medicine between the first patient and the second patient. When the similarity of picking up medicine between the first patient and the second patient exceeds the preset similarity threshold of picking up medicine, the parsing log with an optimized order associated with the running order of picking up medicine for waiting selection is output. When the similarity of picking up medicine between the first patient and the second patient does not exceed the preset similarity threshold of picking up medicine, the parsing log without an optimized order associated with the running order of picking up medicine for waiting selection is output.

[0077] Among them, during the process of the medicine-picking robot picking up medicine from the medicine box in accordance with the running order of picking up medicine for waiting selection, the cloud platform can monitor the first patient of the medicine-picking robot and the second patient of the medicine box. For example, the wearable device, handheld device of the medicine-picking robot or other devices capable of collecting the patient information of the medicine-picking robot transmit the patient information of the medicine-picking robot to the cloud platform, and the medicine box is equipped with a positioning system such as GPS to transmit the patient information of the medicine box to the cloud platform. The preset similarity threshold of picking up medicine can be 30% or 40%. Those skilled in the art can set it based on actual needs.

[0078] For example, as Figure 2 shown, the running order of picking up medicine for waiting selection can be order J1 and order J2. During the process of the medicine-picking robot picking up medicine in accordance with order J1 and order J2, the cloud platform can monitor the first patient of the medicine-picking robot and the second patient of the medicine box. After the medicine-picking robot carries the medicine box and arrives at the second patient 102 picking up medicine from the first patient 101 in accordance with order J1 and completes the medicine-picking for the second patient 102 picking up medicine, the medicine-picking robot can walk through the green channel 201 to pick up medicine for the third patient picking up medicine on the opposite side in accordance with order F1. At this time, the similarity of picking up medicine between the first patient of the medicine-picking robot and the second patient of the medicine box is 50%, exceeding the preset similarity threshold of picking up medicine of 30%. Therefore, the cloud platform outputs the parsing log with an optimized order associated with the running order of picking up medicine for waiting selection.

[0079] It should be noted that as long as the similarity of drug dispensing between the first patient and the second patient checked by the cloud platform exceeds the preset drug dispensing similarity threshold, it is considered to have an optimized order.

[0080] Next, after the drug dispensing robot completes the drug dispensing for the third drug-taking patient 103, it returns to the medicine cabinet according to the order F2, and then continues to carry the medicine cabinet to the fourth drug-taking patient 104. When dispensing drugs for the fourth drug-taking patient 104, the drug dispensing robot never stays far away from the medicine cabinet. For example, the similarity of drug dispensing between the third drug-taking patient 103 and the fourth drug-taking patient 104 is 10%, which does not exceed the preset drug dispensing similarity threshold of 30%. It can be seen that there is no optimized order during the drug dispensing process for the fourth drug-taking patient 104, and there is an optimized order during the drug dispensing process for the third drug-taking patient 103.

[0081] Among them, in this embodiment, the calculation method of the drug dispensing similarity is to segment the traditional Chinese medicine prescription string in the drug dispensing prescription to form a short string vector, obtain the degree of mutual correlation between the short string vectors, and calculate the prescription similarity according to the degree of mutual correlation between the short string vectors.

[0082] For example, the drug dispensing prescription of the third drug-taking patient includes Codonopsis pilosula, Gardenia jasminoides, Bupleurum chinense, Aurantii Fructus Immaturus, Biota orientalis, Atractylodes macrocephala, Paeonia suffruticosa, Leonurus japonicus, Paeonia lactiflora, and Typha angustifolia, and the drug dispensing prescription of the fourth drug-taking patient includes Glycyrrhiza uralensis, Angelica sinensis, Rubia cordifolia, Astragalus membranaceus, Codonopsis pilosula, Ligusticum wallichii, Curcuma aromatica, Scorpio, Mirabilite, and Camphor. Then the similarity of drug dispensing between the third drug-taking patient 103 and the fourth drug-taking patient 104 is 10% (only Codonopsis pilosula is the same), which does not exceed the preset drug dispensing similarity threshold of 30%. It can be seen that there is no optimized order during the drug dispensing process for the fourth drug-taking patient 104, and there is an optimized order during the drug dispensing process for the third drug-taking patient 103.

[0083] For example, when the first patient is a rhinitis patient, his prescription includes Budesonide Nasal Spray, Loratadine, Tranilast, Oxymetazoline Hydrochloride Nasal Drops, and Benralizumab Nasal Spray;

[0084] When the second patient is a sinusitis patient, his prescription includes Budesonide Nasal Spray, Cetirizine, Azelastine Nasal Spray, Tranilast, and Oxymetazoline Hydrochloride Nasal Drops;

[0085] The similarity of drug dispensing between the first patient and the second patient in the medicine cabinet is 60% (Budesonide Nasal Spray, Tranilast, Oxymetazoline Hydrochloride Nasal Drops), which exceeds the preset drug dispensing similarity threshold of 30%. Furthermore, the cloud platform outputs an analysis log with an optimized order associated with the candidate drug dispensing operation order.

[0086] S305. The cloud platform generates multiple drug dispensing orders associated with drug-taking patients based on the optimized order and the candidate drug dispensing operation order.

[0087] After parsing the log with an optimized order associated with the candidate drug dispensing operation sequence, multiple dispensing sequences associated with the dispensing patients can be generated based on the optimized order and the candidate drug dispensing operation sequence. Among them, the optimized order can be a non-drug dispensing operation sequence.

[0088] In some examples, after the cloud platform determines an optimized order associated with the candidate drug dispensing operation sequence, it can determine the optimized order POI, such as the second dispensing patient 102 mentioned above. The staff can go to the second dispensing patient 102, record the order information of the optimized order, and store it in the cloud platform for subsequent generation of the dispensing sequence. It is also possible to create guiding information for the optimized order and store it in the cloud platform together.

[0089] In some embodiments, the cloud platform can also transmit verification information to the processor, which is used to instruct the drug dispensing robot to verify the parsing log. For example, verify whether there is an optimized order. After receiving the verification information, the processor can provide a verification interface to the drug dispensing robot. The drug dispensing robot can enter "has an optimized order" or "does not have an optimized order" in the verification interface, and the processor will feedback the verification log of the drug dispensing robot to the cloud platform.

[0090] After the cloud platform receives the verification log feedback by the processor, it can comprehensively analyze the parsing log and the verification log to determine whether there is an optimized order, and then decide whether to remind the staff to go to the site for investigation. Exemplarily, if there is, go; if not, do not go. When both the parsing log and the verification log indicate that there is an optimized order associated with the candidate drug dispensing operation sequence, multiple dispensing sequences associated with the dispensing patients are generated based on the optimized order and the candidate drug dispensing operation sequence. Through the double-verification method, the rigor of optimizing the dispensing order is increased, and mis-optimization is avoided.

[0091] S306. When the drug dispensing robot dispenses drugs to the dispensing patient, it performs identity verification;

[0092] When the drug dispensing robot stops in front of the dispensing patient, it scans the facial information of the dispensing patient;

[0093] The drug dispensing robot compares the facial information with the information in the patient database to obtain the drug prescription information of the corresponding patient;

[0094] The dispensing patient obtains the dispensed drugs from the drug dispensing robot by himself / herself. When the drug dispensing robot detects that the dispensed drugs obtained by the dispensing patient are inconsistent with the drug prescription information of the corresponding patient, it issues a warning to the cloud platform;

[0095] The drug dispensing robot performs recovery processing on the dispensed drugs.

[0096] S307. The cloud platform uses the selected dispensing operation sequence as the dispensing operation sequence associated with multiple patients picking up medicine.

[0097] In the case where the cloud platform parses that there is no optimized sequence associated with the selected dispensing operation sequence, it indicates that the selected dispensing operation sequence is better than the historical dispensing operation sequence and the dispensing time is shorter. Therefore, the selected dispensing operation sequence can be used as the dispensing operation sequence associated with multiple patients picking up medicine. Subsequently, the dispensing robot picks up medicine according to the selected dispensing operation sequence, which can reduce the dispensing time and improve the dispensing efficiency.

[0098] S308. The cloud platform pushes guidance information to the processor based on the dispensing sequence.

[0099] After outputting the dispensing sequence, the cloud platform can push guidance information to the processor based on the dispensing sequence. Among them, the guidance information includes guidance information for carrying the medicine box and guidance information for not carrying the medicine box. Among them, the guidance information for carrying the medicine box is generated by the selected dispensing operation sequence, and the guidance information for not carrying the medicine box is generated by the optimized sequence.

[0100] S309. The processor provides animation information associated with the optimized sequence to the dispensing robot based on the guidance information.

[0101] After receiving the guidance information, the processor can provide animation information associated with the optimized sequence to the dispensing robot. The animation information can be videos, pictures and texts, etc. The animation information is used to introduce the optimized sequence and how to pick up medicine according to the optimized sequence. After seeing the animation information, the dispensing robot can pick up the goods according to the optimized sequence, thereby reducing the dispensing time and improving the dispensing efficiency.

[0102] Based on the above description, the present application provides an intelligent medicine-taking recognition method based on machine vision. This method can be executed by a cloud platform. In this method, the cloud platform collects medicine-taking data, which includes multiple medicine-taking patients, and collects the actual medicine-taking operation sequence, actual medicine-taking cycle, historical medicine-taking operation sequence, and historical medicine-taking cycle of the medicine-taking robot. Then, the actual medicine-taking operation sequence can be compared with the historical medicine-taking operation sequence first, and the sequences with a deviation value not less than the preset deviation value threshold are output. Then, from these sequences, the sequences with an actual medicine-taking cycle less than the historical medicine-taking cycle are filtered out, that is, the candidate medicine-taking operation sequences are output. Of course, it is also possible to filter out the sequences with a short cycle first and then filter out the sequences with a large deviation value, and then output the candidate medicine-taking operation sequences. After the cloud platform determines the candidate medicine-taking operation sequences, it analyzes whether there is an optimized sequence associated with the candidate medicine-taking operation sequences. When there is, the medicine-taking sequence associated with these multiple medicine-taking patients is generated based on the optimized sequence and the candidate medicine-taking operation sequences. Therefore, subsequent medicine-taking personnel can take medicine for these multiple patients according to this medicine-taking sequence, that is, take medicine by means of the medicine-taking operation medicine box according to the actual medicine-taking operation sequence, so as to reduce the medicine-taking cycle of carrying the medicine box, and take medicine by means of non-carrying the medicine box according to the optimized sequence, so as to reduce the overall medicine-taking cycle. It can be seen that this method can reduce the medicine-taking cycle and improve the medicine-taking efficiency.

[0103] As described above in conjunction with Figures 1 to 3 the intelligent medicine-taking recognition method provided by the embodiments of the present application has been introduced in detail. Next, the systems and devices provided by the embodiments of the present application will be introduced in conjunction with the accompanying drawings.

[0104] As Figure 4 shown, this figure is a schematic diagram of an intelligent medicine-taking recognition system based on machine vision provided by an embodiment of the present application. The system includes:

[0105] A collection unit 401, configured to collect medicine-taking data, where the medicine-taking data includes multiple medicine-taking patients, and collect the actual medicine-taking operation sequence of the medicine-taking robot, the actual medicine-taking cycle of the actual medicine-taking operation sequence, the historical medicine-taking operation sequence associated with the medicine-taking data, and the historical medicine-taking cycle of the historical medicine-taking operation sequence;

[0106] A filtering unit 402, configured to filter out candidate medicine-taking operation sequences from the actual medicine-taking operation sequences, where the order deviation value between the candidate medicine-taking operation sequences and the historical medicine-taking operation sequences is not less than a preset deviation value threshold, and the candidate medicine-taking cycle of the candidate medicine-taking operation sequences is less than the historical medicine-taking cycle;

[0107] An analysis unit 403, configured to analyze whether there is an optimized sequence associated with the candidate medicine-taking operation sequences, and output an analysis log;

[0108] An optimization unit 404, configured to generate a medication-taking order associated with the multiple medication-taking patients based on the optimization order and the candidate medication-taking operation order when the parsed log representation has an optimization order associated with the candidate medication-taking operation order.

[0109] In some possible implementation manners, the parsing unit 403 is specifically configured to output a parsed log with an optimization order associated with the candidate medication-taking operation order when there is a preset patient in the preset ward where the POI is located in the candidate medication-taking operation order; and output a parsed log without an optimization order associated with the candidate medication-taking operation order when there is no preset patient in the preset ward where the POI is located in the candidate medication-taking operation order.

[0110] In some possible implementation manners, the parsing unit 403 is specifically configured to collect a first patient of the medication-taking robot and a second patient of the medication-taking box during the process that the medication-taking box of the medication-taking robot takes medications according to the candidate medication-taking operation order; and output a parsed log with an optimization order associated with the candidate medication-taking operation order when the medication-taking similarity between the first patient and the second patient exceeds a preset medication-taking similarity threshold; and output a parsed log without an optimization order associated with the candidate medication-taking operation order when the medication-taking similarity between the first patient and the second patient does not exceed the preset medication-taking similarity threshold.

[0111] In some possible implementation manners, the optimization unit 404 is further configured to use the candidate medication-taking operation order as the medication-taking operation order associated with the multiple medication-taking patients when the parsed log representation does not have an optimization order associated with the candidate medication-taking operation order.

[0112] In some possible implementation manners, the optimization unit 404 is further configured to push guidance information to the processor based on the medication-taking order, so that the processor provides animation information associated with the optimization order to the medication-taking robot based on the guidance information.

[0113] In some possible implementation manners, the system further includes a transmission unit, configured to transmit verification information to the processor after outputting a parsed log with an optimization order associated with the candidate medication-taking operation order, so that the processor instructs the medication-taking robot to verify the parsed log based on the verification information and feedback a verification log to the cloud platform; the optimization unit 404 is specifically configured to generate a medication-taking order associated with the multiple medication-taking patients based on the optimization order and the candidate medication-taking operation order when both the parsed log and the verification log represent an optimization order associated with the candidate medication-taking operation order.

[0114] In some possible implementation manners, the pre-set patients include patients with phlegm-turbidity obstruction, patients with blood stasis in the heart vessels, patients with qi stagnation in the chest, or patients with cold congealing the heart vessels.

[0115] The machine vision-based intelligent medicine-taking recognition system according to the embodiments of the present application can be associated with the execution of the methods described in the embodiments of the present application, and the above other operations and / or functions of each unit / unit of the machine vision-based intelligent medicine-taking recognition system are for the purpose of implementing Figure 3 the corresponding processes of the various methods in the illustrated embodiments. For the sake of brevity, they will not be described in detail here.

[0116] The embodiments of the present application also provide a computing device. The computing device can be a cloud platform, and the cloud platform is specifically used to implement the functions of the machine vision-based intelligent medicine-taking recognition system as shown in Figure 4 the illustrated embodiments.

[0117] In several embodiments provided by the present application, it should be understood that the disclosed systems, systems, and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the division of units is only a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection can be through some interfaces. The indirect coupling or communication connection of the system or unit can be in an electrical, mechanical, or other form.

[0118] In addition, each functional unit in the various embodiments of the present application can be integrated in a processing unit, or each unit can be physically separate, or two or more units can be integrated in one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit. The above is only the implementation manner of the present application and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the content of the specification and drawings of the present application, or directly or indirectly applied in other related technical fields, are equally included in the patent protection scope of the present application.

[0119] The specific implementation manners of the invention have been described in detail above, but they are only examples. The present application is not limited to the specific implementation manners described above. For those skilled in the art, any equivalent modification or substitution of the invention is also within the scope of the present application. Therefore, equivalent transformations, modifications, improvements, etc. made without departing from the spirit and principle of the present application should all be covered within the scope of the present application.

Claims

1. An intelligent medicine-taking recognition method based on machine vision, characterized in that, Applied to a cloud platform, the method includes: The cloud platform collects dispensing data, and the dispensing data includes multiple dispensing patients; The cloud platform collects the actual dispensing operation sequence of the dispensing robot, the actual dispensing cycle of the actual dispensing operation sequence, the historical dispensing operation sequence associated with the dispensing data, and the historical dispensing cycle of the historical dispensing operation sequence; The actual dispensing operation sequence refers to the sequence in which the dispensing robot advances while carrying the medicine dispensing box during the process of dispensing medicine to the above-mentioned multiple dispensing patients. The actual dispensing cycle refers to the cycle required for the dispensing robot to dispense medicine to multiple dispensing patients according to the actual dispensing operation sequence. The historical dispensing operation sequence refers to the sequence in which the historical dispensing robot advances while carrying the medicine dispensing box during the process of dispensing medicine to the above-mentioned multiple dispensing patients in a historical period. The historical dispensing cycle refers to the cycle required for the historical dispensing robot to dispense medicine to multiple dispensing patients according to the historical dispensing operation sequence. When there are multiple historical dispensing operation sequences, select the sequence with the shortest historical dispensing cycle from the multiple historical dispensing operation sequences; The cloud platform filters out candidate dispensing operation sequences from the actual dispensing operation sequence. The deviation value between the candidate dispensing operation sequence and the historical dispensing operation sequence is not less than a preset deviation value threshold, and the candidate dispensing cycle of the candidate dispensing operation sequence is less than the historical dispensing cycle; The cloud platform analyzes whether there is an optimized sequence associated with the candidate dispensing operation sequence and outputs an analysis log; The cloud platform analyzes whether there are preset patients in the preset ward where the POI is located in the candidate dispensing operation sequence. When there are preset patients in the preset ward where the POI is located in the candidate dispensing operation sequence, output an analysis log indicating that there is an optimized sequence associated with the candidate dispensing operation sequence. When there are no preset patients in the preset ward where the POI is located in the candidate dispensing operation sequence, output an analysis log indicating that there is no optimized sequence associated with the candidate dispensing operation sequence; When the analysis log indicates that there is an optimized sequence associated with the candidate dispensing operation sequence, the cloud platform generates a dispensing sequence associated with the multiple dispensing patients based on the optimized sequence and the candidate dispensing operation sequence; When the dispensing robot dispenses medicine to the dispensing patients, it performs identity verification.

2. The intelligent drug-taking recognition method based on machine vision according to claim 1, wherein, The analyzing whether there is an optimized sequence associated with the candidate dispensing operation sequence and outputting an analysis log includes: Collecting the first patient of the dispensing robot and the second patient of the medicine dispensing box during the process of the medicine dispensing box of the dispensing robot dispensing medicine according to the candidate dispensing operation sequence; When the dispensing similarity between the first patient and the second patient exceeds a preset dispensing similarity threshold, output an analysis log indicating that there is an optimized sequence associated with the candidate dispensing operation sequence; When the dispensing similarity between the first patient and the second patient does not exceed the preset dispensing similarity threshold, output an analysis log indicating that there is no optimized sequence associated with the candidate dispensing operation sequence.

3. The intelligent drug-taking recognition method based on machine vision according to claim 1, characterized in that The method further includes: When the parsed log representation does not have an optimized order associated with the candidate dispensing operation order, the candidate dispensing operation order is used as the dispensing operation order associated with the multiple dispensing patients.

4. The intelligent drug-taking recognition method based on machine vision according to claim 1, characterized in that, The method further includes: Based on the dispensing order, push guiding information to the processor so that the processor provides animation information associated with the optimized order to the dispensing robot based on the guiding information.

5. The intelligent medicine-taking recognition method based on machine vision according to claim 2 or 3, characterized in that, After outputting the parsed log with an optimized order associated with the candidate dispensing operation order, the method further includes: transmitting verification information to the processor so that the processor instructs the dispensing robot to verify the parsed log based on the verification information and feedback the verification log to the cloud platform; When the parsed log representation has an optimized order associated with the candidate dispensing operation order, the cloud platform generates the dispensing order associated with the multiple dispensing patients based on the optimized order and the candidate dispensing operation order, including: When both the parsed log and the verification log represent having an optimized order associated with the candidate dispensing operation order, generate the dispensing order associated with the multiple dispensing patients based on the optimized order and the candidate dispensing operation order.

6. The method according to claim 2, wherein When the preset patient is a traditional Chinese medicine patient, it includes at least: Patients with phlegm turbidity blocking, patients with blood stasis in the heart vessels, patients with qi stagnation in the chest, or patients with cold congealing the heart vessels; When the preset patient is a Western medicine patient, it includes at least: patients with rhinitis, patients with sinusitis, patients with pharyngitis.

7. The method according to claim 1, characterized in that, When the dispensing robot dispenses medicine to the dispensing patient and conducts identity verification, it further includes: When the dispensing robot stops in front of the dispensing patient, scan the facial information of the dispensing patient; The dispensing robot compares the facial information with the information in the patient database to obtain the drug prescription information of the corresponding patient; When the dispensing patient obtains the dispensed medicine by himself / herself from the dispensing robot, when the dispensing robot detects that the dispensed medicine obtained by the dispensing patient is inconsistent with the drug prescription information of the corresponding patient, send a warning to the cloud platform; The dispensing robot conducts recycling processing on the dispensed medicine.

8. An intelligent medicine-taking recognition system based on machine vision, characterized in that, For implementing an intelligent dispensing recognition method based on machine vision as described in any one of claims 1-7, the system includes: An acquisition unit for acquiring dispensing data, where the dispensing data includes multiple dispensing patients, the actual dispensing operation order of the dispensing robot, the actual dispensing cycle of the actual dispensing operation order, the historical dispensing operation order associated with the dispensing data, and the historical dispensing cycle of the historical dispensing operation order; A filtering unit for filtering out candidate dispensing operation orders from the actual dispensing operation order, where the order deviation value of the candidate dispensing operation order from the historical dispensing operation order is not less than a preset deviation threshold value, and the candidate dispensing cycle of the candidate dispensing operation order is less than the historical dispensing cycle; An analysis unit for analyzing whether there is an optimized order associated with the candidate dispensing operation order and outputting an analysis log; Optimization unit, configured to generate a dispensing order associated with the multiple dispensing patients based on the optimization order and the candidate dispensing operation order when the parsed log representation has an optimization order associated with the candidate dispensing operation order.

Images