Cash register abnormity alarm method and system, and cash register

Abstract

The invention discloses a cash register abnormity alarm method and system and a cash register, and the method comprises the steps: S1, obtaining the operation data of the cash register during the execution of a task, judging whether the operation data is abnormal, and generating abnormity detection data if the operation data is abnormal; s2, based on a preset intelligent identification model, analyzing the anomaly detection data, judging an anomaly type and determining a corresponding alarm scheme; and S3, dynamically triggering a multi-mode cooperative alarm mechanism based on the corresponding alarm scheme, and providing an intelligent processing suggestion. The reliability and effectiveness of the alarm are improved, and the difficulty of the cashier to handle the abnormity is reduced due to the provision of the intelligent processing suggestions.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cash registers, in particular to a cash register abnormal alarm method, system and cash register. BACKGROUND

[0002] A cash register is an indispensable core equipment in the retail industry, mainly used for handling commodity sales and payment-related transactions. Traditional cash registers have basic functions such as commodity scanning, price calculation, and payment processing. With the development of information technology, modern cash registers have further integrated advanced functions such as inventory management and sales data analysis. Inventory management functions can update real-time commodity inventory information, helping businesses to grasp inventory status and optimize replenishment plans; sales data analysis functions reveal sales trends and customer purchase preferences by analyzing and mining sales data, providing a basis for business decision-making. The introduction of these functions has significantly improved the business management capabilities of businesses.

[0003] However, existing cash registers still have obvious deficiencies in abnormal alarm. Currently, the abnormal alarm of most cash registers is limited to detecting hardware failures (such as scanner malfunction) or software errors (such as program crashes), and lacks effective automatic monitoring and processing means for more complex business abnormalities, such as inventory data inconsistency (such as actual inventory and records not matching) or sales data abnormalities (such as abnormal return rate). Although modern cash registers can collect a large amount of data, their data analysis capabilities are limited, making it difficult to achieve in-depth mining and real-time processing, and unable to timely discover hidden abnormal situations. For example, when the bar code of a commodity is damaged and cannot be recognized by the scanner in a busy supermarket, the cash register only displays a simple prompt and fails to take further measures in time, and the cashier needs to repeat the attempt or manually operate, interrupting the transaction process and affecting efficiency. At the same time, the single alarm mode (such as buzzer or text prompt) is easy to be ignored in a noisy environment, and the system also lacks automatic diagnosis capability, which cannot distinguish the cause of the abnormality (such as damaged bar code or hardware failure), increasing the burden of manual troubleshooting.

[0004] There is an urgent need for a cash register abnormal alarm method, system and cash register. SUMMARY

[0005] The present application provides a cash register abnormal alarm method, system and cash register to solve the above problems in the prior art.

[0006] In order to achieve the above purpose, the present application provides the following technical scheme:

[0007] A cash register abnormal alarm method, comprising:

[0008] S1: obtaining running data of the cash register when performing a task, determining whether the running data has an abnormal situation, and generating abnormal detection data if there is an abnormal situation;

[0009] S2: Analyze the abnormal detection data based on the preset intelligent recognition model, judge the abnormal type and determine the corresponding alarm scheme;

[0010] S3: Based on the corresponding alarm scheme, dynamically trigger the multi-mode cooperative alarm mechanism, and provide intelligent processing suggestions.

[0011] Among them, S1 step includes:

[0012] S11: Real-time collection of cash register running data, including scanner status, payment system status and network connection status;

[0013] S12: Compare the collected running data with the preset normal running threshold to identify abnormal data points;

[0014] S13: According to the distribution characteristics of abnormal data points, generate abnormal detection data.

[0015] Among them, S2 step includes:

[0016] S21: Input the abnormal detection data into the intelligent recognition model to extract the abnormal feature vector;

[0017] S22: Based on the abnormal feature vector, match in the preset abnormal type library to determine the abnormal type;

[0018] S23: According to the abnormal type and the current cash register environment, select the corresponding alarm scheme from the preset alarm scheme library.

[0019] Among them, S3 step includes:

[0020] S31: According to the corresponding alarm scheme, determine the alarm trigger condition and alarm mode combination;

[0021] S32: Real-time monitoring of alarm trigger conditions, when the conditions are met, trigger multiple alarm modes synchronously;

[0022] S33: Based on the abnormal type and the current cash register state, generate intelligent processing suggestions and push to the cashier interface.

[0023] Among them, S12 step includes:

[0024] S121: Establish a time series model of normal running data;

[0025] S122: Calculate the deviation value of the current running data and the time series model;

[0026] S123: When the deviation value exceeds the preset threshold, mark it as an abnormal data point.

[0027] Among them, S22 step includes:

[0028] S221: constructing a multi-dimensional feature space of the abnormal feature vector;

[0029] S222: calculating the similarity between the abnormal feature vector and each abnormal type feature vector;

[0030] S223: selecting the abnormal type with the highest similarity as the final judgment result.

[0031] The S32 step includes:

[0032] S321: determining the trigger sequence of the alarm mode according to the alarm scheme, the alarm mode including sound, light, screen display, and vibration;

[0033] S322: monitoring the execution state of each alarm mode in real time;

[0034] S323: when detecting that a certain alarm mode is executed abnormally, automatically switching to a backup alarm mode.

[0035] The cash register abnormal alarm system includes:

[0036] A data monitoring module is configured to obtain running data of the cash register when performing a task, determine whether the running data has an abnormal condition, and generate abnormal detection data if the abnormal condition exists;

[0037] An intelligent analysis module is configured to analyze the abnormal detection data based on a preset intelligent recognition model, determine an abnormal type, and determine a corresponding alarm scheme;

[0038] An alarm execution module is configured to dynamically trigger a multi-mode cooperative alarm mechanism based on the corresponding alarm scheme and provide intelligent processing suggestions.

[0039] The cash register includes:

[0040] An abnormal alarm system is configured to execute the cash register abnormal alarm method;

[0041] A main control unit is connected to the abnormal alarm system and is configured to perform cash register operations and provide running state data to the abnormal alarm system;

[0042] A barcode scanner is configured to scan a product barcode and transmit data to the main control unit, and the abnormal alarm system is configured to monitor whether the barcode scanner is scanned abnormally;

[0043] A touch display screen is configured to display cash register information and receive user input, and switch to a specific alarm interface when receiving an alarm signal from the abnormal alarm system;

[0044] A multi-mode alarm device includes a sound alarm, an indicator light, and a vibration motor, and is configured to realize multi-mode cooperative alarm according to the control of the abnormal alarm system.

[0045] Further comprising:

[0046] A cashbox is used to store cash and is connected with the abnormal alarm system to monitor the opening abnormality.

[0047] A printer is used to print the cash register receipt and is connected with the abnormal alarm system to monitor the printing abnormality.

[0048] Compared with the prior art, the present application has the following advantages:

[0049] A cash register abnormal alarm method, system and cash register, the abnormal alarm method comprising: S1: obtaining the running data of the cash register when performing a task, judging whether the running data has an abnormal condition, and generating abnormal detection data if there is an abnormal condition; S2: based on the preset intelligent recognition model, analyzing the abnormal detection data, judging the abnormal type and determining the corresponding alarm scheme; S3: based on the corresponding alarm scheme, dynamically triggering a multi-mode cooperative alarm mechanism, and providing intelligent processing suggestions. The reliability and effectiveness of the alarm are improved, and the difficulty of the cashier in processing the abnormality is reduced by providing intelligent processing suggestions.

[0050] Other features and advantages of the present application will be described in the following description, and some will become apparent from the description, or will be understood by those skilled in the art through implementation of the present application.

[0051] The technical solutions of the present application will be further described in detail below with the help of the drawings and examples. BRIEF DESCRIPTION OF DRAWINGS

[0052] The accompanying drawings are used to provide a further understanding of the present application, and constitute a part of the specification, together with the embodiments of the present application, to explain the present application, and do not constitute a limitation on the present application. In the drawings:

[0053] Figure 1 A flowchart of a cash register abnormal alarm method in an embodiment of the present application;

[0054] Figure 2 A flowchart of analyzing abnormal detection data in an embodiment of the present application;

[0055] Figure 3 A structural diagram of a cash register abnormal alarm system in an embodiment of the present application. DETAILED DESCRIPTION

[0056] The preferred embodiments of the present application will be described below with reference to the accompanying drawings, and it should be understood that the preferred embodiments described herein are only used to illustrate and explain the present application, and do not limit the present application.

[0057] The embodiments of the present application provide a cash register abnormal alarm method as Figure 1As shown, a cash register abnormal alarm method comprises:

[0058] S1: Obtain the running data of the cash register when performing tasks, judge whether the running data has abnormal conditions, and generate abnormal detection data if there is an abnormal condition;

[0059] S2: Based on the preset intelligent recognition model, analyze the abnormal detection data, judge the abnormal type and determine the corresponding alarm scheme;

[0060] S3: Based on the corresponding alarm scheme, dynamically trigger the multi-mode cooperative alarm mechanism, and provide intelligent processing suggestions.

[0061] The working principle of the above technical solution is: S1: Obtain the running data of the cash register when performing tasks, judge whether the running data has abnormal conditions, and generate abnormal detection data if there is an abnormal condition; wherein, the cash register running data is obtained by a plurality of sensors configured inside the cash register; the running data includes cash register hardware running state data, transaction data and operation behavior data; the hardware running state data mainly includes: cash register temperature, voltage, money box switch state, printer paper quantity, etc.; the transaction data includes: single transaction amount, transaction frequency, transaction time, discount application situation, etc.; the operation behavior data includes: operator login information, opening and closing time, invalid operation times, key response time, etc. Abnormal judgment is based on the preset normal running parameter range, when one or more running data exceeds the corresponding normal range, the system will automatically generate abnormal detection data.

[0062] S2: Based on the preset intelligent recognition model, analyze the abnormal detection data, judge the abnormal type and determine the corresponding alarm scheme; wherein, the intelligent recognition model is trained through a large number of historical abnormal cases, which can accurately distinguish different types of abnormal conditions. The abnormal type includes: hardware fault type exception, operation error type exception, suspicious transaction type exception and system performance type exception. Hardware fault type exception refers to the abnormal work of device components; operation error type exception refers to the problem caused by the error operation of the cashier; suspicious transaction type exception refers to the financial risk behavior; system performance type exception refers to slow software running or abnormal response, etc. Each type of exception corresponds to a specific alarm scheme, which includes alarm level, alarm receiving object, alarm method and processing time limit. The alarm level is divided into slight prompt, general warning, serious warning and emergency warning; the alarm receiving object can be a cashier, a store manager, a technical support personnel or a financial management personnel; the alarm method can be screen prompt, sound alarm, short message notification or system lock, etc.

[0063] S3: Based on the corresponding alarm scheme, dynamically trigger the multi-mode cooperative alarm mechanism and provide intelligent processing suggestions; wherein the multi-mode cooperative alarm mechanism refers to activating multiple alarm channels simultaneously or in sequence according to the severity and type of different abnormal situations. For example, for minor operation errors, only display prompt information on the screen; while for suspicious large transaction abnormalities, sound alarms, lock the cash register, and send a text message to the store manager. Intelligent processing suggestions are automatically generated based on historical successful solutions, providing corresponding processing steps and suggestions for different abnormal types. For example: if the paper quantity is insufficient, prompt the specific operation steps to replenish the printer paper; for suspicious transaction alarms, provide inspection procedures and temporary handling measures.

[0064] Also included are:

[0065] Based on the cash register operating environment and business peak characteristics, the alarm triggering sensitivity is adjusted;

[0066] According to the alarm triggering sensitivity adjustment, the frequency and judgment threshold of abnormal detection are configured.

[0067] The beneficial effects of the above technical solutions are: by analyzing the operating environment and business peak characteristics of the cash register, the sensitivity of alarm triggering is dynamically adjusted, for example: during the business peak period, the detection frequency of some non-critical parameters is reduced to ensure system response speed; while during the night inventory period, the detection sensitivity of financial abnormalities is increased; this dynamic adjustment mechanism can ensure timely detection of key abnormalities and avoid resource waste and frequent false alarms.

[0068] In another embodiment, the S1 step includes:

[0069] S11: Real-time acquisition of cash register operating data, including scanner status, payment system status and network connection status;

[0070] S12: Compare the collected operating data with the preset normal operating threshold to identify abnormal data points;

[0071] S13: Generate abnormal detection data according to the distribution characteristics of the abnormal data points.

[0072] The working principle of the above technical solution is as follows: S11: real-time collection of operation data of the cash register to determine the current cash register working state. The operation data is collected in real time through sensors and monitoring programs configured on each functional module of the cash register. The collected operation data mainly includes scanner state, payment system state and network connection state. The scanner state monitoring is realized through the state detection module built-in the bar code scanner. The success rate is calculated by the ratio of the number of successful scans to the total number of scan attempts in a preset time window. When the ratio is less than 90%, it is determined to be an abnormal state. The response time is obtained by measuring the time interval from triggering the scan signal to outputting the barcode data. The normal range is set to 50-200 milliseconds. If it exceeds this range, it is marked as a response anomaly. The light source intensity is monitored in real time by the photosensitive sensor integrated in the scanner. The monitoring data is compared with the preset standard light intensity value. If the deviation exceeds ±15%, the light source anomaly alarm is triggered. The payment system state monitoring includes real-time detection of the bank card reader and mobile payment interface. The card reader state is judged by periodically sending a standard test signal and detecting the return response to judge the hardware connection integrity. The response time and success rate of card transaction are also counted. If the response time exceeds 3 seconds or the success rate is less than 95%, it is determined to be a system anomaly. The mobile payment interface state is monitored through the heartbeat detection mechanism established with the third-party payment platform. A status query request is sent every 30 seconds. If there is no response for 3 consecutive times, it is determined to be a connection anomaly, and the automatic reconnection mechanism is started. The network connection state monitoring realizes the connection stability evaluation through multiple detection mechanisms. The system periodically sends ping test packets to the background server to count the packet loss rate and connection interruption times. If the packet loss rate exceeds 5% or the interruption times in a single hour exceed 3 times, it is marked as unstable connection. The data transmission rate is measured by sending test data packets of a fixed size and calculating the transmission time. When the transmission rate is lower than the set minimum bandwidth requirement, the network performance alarm is triggered. The network delay is obtained by measuring the round-trip time (RTT) of the data packet. The normal delay range is set to 50-500 milliseconds. If it exceeds the range, the system automatically optimizes the network parameters or switches to the standby network connection. The abnormal data recognition adopts a multi-parameter comprehensive judgment mechanism. When any monitoring parameter exceeds the normal threshold range, the system immediately records the abnormal event and generates a status report. For minor abnormalities, the system automatically performs repair operations such as restarting related modules or adjusting parameter settings. For serious abnormalities, the system issues an audible and visual alarm and notifies the management personnel for manual intervention.

[0073] S12: The collected operational data is compared with preset normal operating thresholds to identify abnormal data points. These preset normal operating thresholds are reasonable ranges for various indicators derived from statistical analysis of historical cash register operational data. The comparison process includes: normalizing the collected operational data, comparing the processed data with the corresponding thresholds, and marking any data point exceeding the threshold range as an abnormal data point. Abnormal data point identification also considers the time-series characteristics of the data. By comparing the data change trends across multiple consecutive time points, misjudgments caused by occasional fluctuations can be eliminated, improving the accuracy of anomaly identification.

[0074] S13: Based on the distribution characteristics of abnormal data points, generate abnormal detection data and trigger an alarm. The abnormal detection data includes one or more of the following: abnormal type, abnormal severity, abnormal duration, and abnormal location. By analyzing the distribution pattern of abnormal data points, the system can determine whether the abnormality is a temporary fault or a systemic problem. The abnormal severity is calculated based on the distance from the normal threshold and is used to determine the urgency of the alarm. The abnormal duration is the length of time the abnormality is continuously detected, used to filter out short-term fluctuations. The abnormal location is precisely pinpointed to a specific hardware module or software functional area, facilitating rapid problem location by maintenance personnel.

[0075] The beneficial effects of the above technical solution are as follows: By analyzing the workload of the cash register, the detection frequency can be reduced during peak customer traffic periods to minimize the impact on checkout efficiency; while the detection frequency can be increased when the workload is low to detect potential problems early. Simultaneously, based on historical anomaly records, the system will increase the monitoring density for modules with frequent problems and appropriately reduce detection for modules with high stability, achieving rational resource allocation and accurate anomaly detection.

[0076] In another embodiment, such as Figure 2 As shown, step S2 includes:

[0077] S21: Input the anomaly detection data into the intelligent recognition model and extract the anomaly feature vector;

[0078] S22: Based on the abnormal feature vector, match it in the preset abnormal type library to determine the abnormal type;

[0079] S23: Select the corresponding alarm scheme from the preset alarm scheme library based on the anomaly type and the current POS environment.

[0080] The working principle of the above technical solution is as follows: S21: Input the anomaly detection data into the intelligent recognition model and extract the anomaly feature vector. The anomaly detection data is collected through various sensing devices such as cameras, weight sensors, barcode scanners, and touch screens configured around the cash register. The anomaly detection data can include multi-dimensional information such as the cash register operation screen, the placement status of goods, the cashier's operation behavior, and customer interaction. The intelligent recognition model adopts a deep learning neural network algorithm. First, it preprocesses and extracts features from the collected multi-dimensional data. It extracts visual features from the image data through a convolutional neural network and performs numerical processing and normalization on the sensor data. The model establishes a baseline pattern library of normal cash register operations through training samples and calculates the deviation value between the real-time detection data and the baseline pattern. When the deviation value exceeds a preset threshold, the model identifies it as abnormal behavior and generates a corresponding anomaly feature vector. The intelligent recognition model mainly analyzes whether the cashier's hand movements conform to the standard process, whether the operation speed is abnormal, and whether there are repeated or skipped operations. At the same time, it analyzes transaction data such as the amount of a single item, the matching degree of the product category, the fluctuation of the total transaction amount, and the settlement method. The model ultimately outputs structured identification results that include anomaly type, anomaly severity score, and confidence level, providing a basis for decision-making in subsequent anomaly handling.

[0081] S22: Based on the abnormal feature vector, match it against a pre-defined abnormality type library to determine the abnormality type. The extraction of the abnormal feature vector includes: performing temporal and spatial feature extraction on the collected multi-dimensional information; then, using a pre-trained neural network model to perform dimensionality reduction and feature enhancement on the extracted raw features to obtain representative abnormal feature vectors. The abnormality type matching process includes: calculating the similarity between the extracted abnormal feature vector and the pre-configured abnormality type feature library; and determining the pre-defined type of the current abnormality based on a set similarity threshold. The pre-defined abnormality type library is derived from the analysis of numerous historical cases and includes common abnormal behavior patterns, such as missed product scanning, price substitution, cash shortage, and discrepancies in accounts.

[0082] S23: Based on the anomaly type and the current checkout environment, select the corresponding alarm scheme from the preset alarm scheme library. The alarm scheme includes one or more of the following: alarm levels for different recipients, alarm information transmission methods, and alarm response procedures. The alarm level is primarily determined by the severity and frequency of the anomaly, as well as potential risk assessment. The alarm level determines whether an alert is sent to the frontline supervisor or the security department. Alarm information transmission methods include mobile terminal push notifications, system pop-up alerts, and audible and visual alarms. The alarm response procedure defines the standard handling procedures that personnel in each position should take upon receiving an alarm, ensuring that anomalies are handled promptly and effectively.

[0083] Also includes:

[0084] Based on historical alarm data and processing feedback, the alarm threshold is adaptively adjusted.

[0085] Configure differentiated anomaly detection rules based on peak business hours and special promotional activities.

[0086] The beneficial effects of the above technical solution are as follows: by analyzing and processing historical alarm data, the alarm threshold is continuously optimized, reducing the false alarm rate while ensuring the capture rate of real anomalies. Specific adaptive adjustments can select the most suitable anomaly judgment rules from a pre-configured rule base based on the business characteristics of different areas and time periods within the shopping mall, thereby improving the accuracy and practicality of anomaly detection.

[0087] In another embodiment, step S3 includes:

[0088] S31: Determine the combination of alarm triggering conditions and alarm methods according to the corresponding alarm scheme;

[0089] S32: Real-time monitoring of alarm triggering conditions; when the conditions are met, multiple alarm modes are triggered simultaneously.

[0090] S33: Based on the exception type and the current checkout status, generate intelligent handling suggestions and push them to the cashier's interface.

[0091] The working principle of the above technical solution is as follows: S31: Determine the alarm triggering conditions and alarm method combination according to the corresponding alarm scheme; wherein, the step of determining the alarm triggering conditions includes: matching the characteristics of the current cashier abnormal event from the pre-built abnormal behavior model library to determine the first-level triggering threshold; obtaining the current cashier environment status, analyzing the environment status, and determining the second-level triggering threshold; comprehensively analyzing the first-level triggering threshold and the second-level triggering threshold to determine the final alarm triggering conditions. For example: anomalies can be categorized into minor, moderate, and severe anomalies, each corresponding to different processing priorities; the anomaly behavior model library is a collection of anomaly patterns extracted from historical transaction data using machine learning algorithms, including various anomaly types such as barcode recognition failure, price mismatch, duplicate scanning, and abnormal product weight; when the anomaly is barcode recognition failure, the first-level trigger threshold may be set to moderate anomaly; the environmental status analysis steps are: feature extraction of the checkout environment data, construction of an environmental parameter set based on the extracted feature parameters, and then retrieval of corresponding associated trigger adjustment parameters from a pre-configured environmental impact analysis library based on the environmental parameter set; environmental data includes: current checkout customer traffic, cashier working hours, system load status, store operating hours (peak or off-peak), etc.; the parameters in the environmental parameter set... The data includes parameters representing current customer traffic at the checkout counter, cashier working hours, system load status, and store operating hours. The environmental impact analysis library is pre-configured, associating environmental parameters with trigger adjustment parameters. The steps for determining alarm method combinations include: selecting a suitable alarm method combination from the alarm method resource library based on the final determined alarm trigger conditions; the alarm method resource library stores various alarm notification methods and their applicable scenarios, such as cashier screen prompts, voice reminders, administrator mobile terminal notifications, and light prompts; different intensity and coverage alarm methods are selected according to the severity and urgency of the anomaly; for barcode recognition failure anomalies, the system may select a combination of cashier screen prompts and voice reminders to ensure that cashiers can notice the anomaly in a timely manner and take appropriate measures.

[0092] S32: Real-time monitoring is implemented through: a first monitoring unit installed at the checkout counter; and / or a second monitoring unit built into the checkout system; and / or a third monitoring unit formed by external sensors connected to the checkout equipment. The first monitoring unit includes an image acquisition device, a weight sensor, and a barcode reader installed at the checkout counter. The second monitoring unit includes a transaction data analysis module; the transaction data analysis module includes one or more of the following: a price consistency check module, a product identification accuracy verification module, and an operation timing analysis module. The third monitoring unit includes auxiliary monitoring devices such as shelf monitoring cameras, a customer behavior analysis system, and store environment sensors.

[0093] The data processing steps include: comparing the product images, weights, and barcode information acquired by the first monitoring unit with information in the product database; performing real-time analysis of the data consistency and logic during the transaction process through the second monitoring unit; and constructing a complete checkout context awareness model by combining environmental data provided by the third monitoring unit. The monitoring judgment steps are as follows: when the barcode reader in the first monitoring unit cannot recognize the product barcode or the recognition result does not match expectations, a barcode anomaly judgment is triggered; the system will immediately compare this anomaly with preset alarm trigger conditions; if the trigger conditions are met, the next alarm action will be executed immediately.

[0094] The steps for simultaneously triggering multiple alarm methods include: activating multiple alarm channels simultaneously according to a pre-determined combination of alarm methods; pushing abnormal information to the cashier's work interface, including the type of abnormality, abnormal product information, and possible causes; simultaneously playing a voice reminder through the cashier's speaker to alert the cashier to the current abnormality; and, if necessary, sending a notification to the area administrator's mobile terminal so that the administrator can be informed of the abnormality in a timely manner and assist in handling it. The content of the alarm information will be customized according to the different recipients to ensure that the information is accurate, concise, and easy to understand and execute.

[0095] S33: The steps for generating intelligent processing suggestions include: the system first identifies the anomaly type and retrieves a set of processing solutions corresponding to that type of anomaly from the processing strategy library; it then evaluates the applicability of the processing solution set based on the current checkout status parameters; according to the evaluation results, the processing solutions are ranked according to indicators such as processing efficiency, success rate, and impact on customer experience; finally, intelligent processing suggestions are generated, including a primary suggestion and alternative suggestions. The processing strategy library is a knowledge base formed through historical processing case analysis, expert knowledge extraction, and machine learning algorithm training, containing effective processing methods for various anomaly situations; the checkout status parameters include the current transaction progress, the number of people waiting in the queue, the cashier's proficiency, and the status of available resources.

[0096] Taking barcode recognition failure as an example, the system analyzes possible reasons for the failure, such as barcode damage, printing quality issues, or reader malfunction. For damaged barcodes, the system prioritizes product appearance recognition technology. The steps of product appearance recognition include: capturing multi-angle images of the product using a cashier camera; feeding the captured images into a pre-trained deep learning model for feature extraction; the feature extraction model identifying visual features such as shape, color, and packaging design; comparing the extracted features with a product image database to generate a list of matching products; displaying the most likely product options to the cashier based on matching scores; and the cashier selecting the correct product via a touch interface to complete the recognition process. The deep learning model is a neural network trained on a large number of product image samples, capable of adapting to challenges in real-world scenarios such as varying lighting conditions, partial occlusion, and changing angles.

[0097] The process of pushing processing suggestions includes: the system generating a structured data packet containing exception information and processing suggestions; transmitting the data packet to the display module of the cashier's operation interface through the internal communication mechanism of the POS system; the display module presenting the processing suggestions in an intuitive and eye-catching manner according to the preset interface template; the push interface using visual elements such as color coding and icon prompts to enhance information recognition; for complex processing steps, the system provides step-by-step guidance to reduce the difficulty of operation; the display of processing suggestions takes into account the cashier's operating habits and the actual situation of the store, providing corresponding operation guidance for the current scenario to ensure efficient completion of exception handling and minimize the impact on customer waiting time.

[0098] The beneficial effects of the above technical solution are: by determining the combination of alarm triggering conditions and alarm methods, the alarm system becomes more flexible and controllable, and can adapt to different abnormal situations and cashier environments.

[0099] In another embodiment, step S12 includes:

[0100] S121: Establish a time series model of normal operation data;

[0101] S122: Calculate the deviation between the current running data and the time series model;

[0102] S123: When the deviation value exceeds the preset threshold, it is marked as an abnormal data point.

[0103] The working principle of the above technical solution is as follows: S121: Establish a time series model of normal operation data; normal operation data is obtained through various operational data recorded by the cash register during daily business operations; the basis for selecting these operational data is that they can directly reflect the status and health of the cash register's core business processes; specifically, this includes cashier login time for monitoring system startup and login behavior patterns; transaction amount for tracking changes in core business volume; transaction frequency for measuring business activity; transaction time distribution for identifying business characteristics in different time periods; number of refund operations for monitoring abnormal or problematic transactions; and drawer opening frequency for associating with cash management operations; before establishing the model, historical data needs to be cleaned and standardized; methods for identifying and judging abnormal data include setting thresholds based on statistical distribution. For example, data points exceeding the historical mean plus or minus three standard deviations, or data points below or above the historical percentile (e.g., 1%) or above the historical percentile (e.g., 99%), as well as outliers that clearly do not conform to business logic, should be identified. Simultaneously, business rules should be considered, such as transactions or logins occurring outside of business hours, unusually large or small single transaction amounts, and frequent drawer openings within a short period. These identified noise and outlier data points should be removed. Then, an appropriate time series modeling method should be selected based on business characteristics, such as using the moving average method. After model training, the model should be able to reflect the normal operating parameters of the cash register under different time periods and business scenarios. Based on this model, a corresponding alarm scheme should be implemented. When the real-time monitored operational data indicators continuously or significantly deviate from the normal range predicted by the model, a tiered alarm should be triggered.

[0104] S122: Calculate the deviation between the current operating data and the time series model; where the current operating data refers to the various operational parameters collected in real time by the cash register; when calculating the deviation, the current data first needs to be preprocessed according to the same standard as when modeling; then, the processed current data is compared with the normal value predicted by the time series model; the comparison method can be to calculate Euclidean distance, Mahalanobis distance or other statistical distance measures; the deviation calculation also needs to consider the business cycle factors, such as the different characteristics of weekdays and weekends, promotional periods and weekdays; the calculation result can be a single value or a comprehensive score of multi-dimensional feature vectors;

[0105] S123: When the deviation value exceeds the preset threshold, it is marked as an abnormal data point. The preset threshold can be determined based on historical data analysis by setting a confidence interval; alternatively, different levels of thresholds can be manually set according to business needs, corresponding to different levels of alarms. When the system detects that the deviation value exceeds the preset threshold, it will immediately mark the data point as an abnormal point and trigger the alarm mechanism. The alarm methods can include various forms such as system interface prompts, audible and visual alarms, and administrator SMS notifications. Abnormal data points will be recorded in a dedicated log for subsequent analysis and tracking. The system will also provide preliminary analysis of the cause of the abnormality and handling suggestions based on the type of abnormality.

[0106] Also includes:

[0107] Based on the busyness of the checkout process and the level of abnormal risk, an alarm triggering frequency analysis was conducted.

[0108] Based on the analysis of alarm trigger frequency, the execution interval of anomaly detection is dynamically adjusted.

[0109] The beneficial effects of the above technical solution are as follows: by assessing the level of business activity and risk, the frequency of anomaly detection can be reasonably arranged, ensuring security monitoring while avoiding excessive consumption of system resources. For example, during peak business hours, the detection frequency can be appropriately reduced but the sensitivity increased, while during off-peak hours, more frequent but less sensitive detections can be performed, ensuring both system response speed and timely detection of abnormal behavior.

[0110] In another embodiment, step S22 includes:

[0111] S221: Construct a multidimensional feature space for abnormal feature vectors;

[0112] S222: Calculate the similarity between the anomaly feature vector and the feature vectors of each anomaly type;

[0113] S223: Select the anomaly type with the highest similarity as the final judgment result.

[0114] The working principle of the above technical solution is as follows: S221: Construct a multi-dimensional feature space for abnormal feature vectors; wherein, the construction of abnormal feature vectors requires extracting relevant features from various operational data of the cash register, including: cash register transaction frequency, single transaction amount, business hours distribution, product scanning interval time, product category distribution, cashier operation behavior characteristics, etc.; after these raw data are preprocessed, they are converted into standardized numerical forms; then, the system uses a feature selection algorithm to select the feature dimensions that are most distinguishable for anomaly identification; finally, these selected feature dimensions are combined to form a multi-dimensional feature space, in which each abnormal event can be represented as a feature vector; this feature space needs to have sufficient dimensions to capture the differential features of various abnormal patterns;

[0115] S222: Calculate the similarity between the anomaly feature vector and the feature vectors of each anomaly type. The anomaly type feature vectors are pre-established based on historical anomaly data, including various anomaly types such as counterfeit currency use, fraudulent product scanning, cashier embezzlement, system failure, malicious refunds, and abnormal transaction cancellations. The system extracts standard feature vectors for each type from the anomaly feature library. Then, similarity calculation methods, such as cosine similarity, Euclidean distance, and Mahalanobis distance, are used to calculate the similarity between the currently detected anomaly feature vector and the feature vectors of each anomaly type. The similarity calculation considers the weights of different feature dimensions, with important features receiving higher weight values. The higher the similarity, the higher the match between the current anomaly and that type of anomaly.

[0116] S223: Select the anomaly type with the highest similarity as the final judgment result; the system will sort all similarity values ​​calculated in the previous step and find the anomaly type with the highest similarity; if the highest similarity exceeds a preset threshold, the system will confirm the anomaly type as the final judgment result; otherwise, it may be marked as "unknown anomaly type"; the final judgment result will trigger the corresponding alarm mechanism, including screen prompts, sound alarms, administrator notifications, etc.; at the same time, the system will record the detailed information of this anomaly event, including the occurrence time, anomaly characteristics, judgment type, etc., for subsequent analysis and system optimization; for specific types of anomalies that recur repeatedly, the system will also adaptively adjust the feature vector of that type to improve the accuracy of future recognition.

[0117] The beneficial effects of the above technical solution are as follows: by constructing a multi-dimensional feature space and comparing similarity, it enables accurate identification and classification of abnormal situations in cash registers, thereby helping merchants effectively prevent various cash register risks and improve operational security. The system will periodically update the feature space and various types of feature vectors based on newly added abnormal samples to maintain the timeliness and accuracy of the alarm mechanism.

[0118] In another embodiment, step S32 includes:

[0119] S321: Determine the triggering sequence of alarm methods according to the alarm scheme. Alarm methods include sound, light, screen display and vibration.

[0120] S322: Real-time monitoring of the execution status of each alarm mode;

[0121] S323: When an abnormality is detected in the execution of a certain alarm mode, the system will automatically switch to the backup alarm mode.

[0122] The working principle of the above technical solution is as follows: S321: Determine the triggering order of alarm methods according to the alarm scheme; wherein, the alarm scheme can be set and adjusted through the operation interface of the cash register; the alarm methods include sound alarm, light alarm, screen display alarm and vibration alarm; the determination of the triggering order is based on the actual needs of the cash register scenario, and sound alarm can be the first choice alarm method, light alarm as the second choice alarm method, and screen display alarm and vibration alarm as alternative alarm methods; the alarm scheme settings also include specific parameters for each alarm method, such as the volume of the sound alarm, the flashing frequency of the light alarm, the prompt content of the screen display alarm, the vibration intensity of the vibration alarm, etc.

[0123] S322: Real-time monitoring of the execution status of each alarm mode; the monitoring of the execution status is achieved through the sensors and detection modules built into the cash register; the execution status of the sound alarm is detected by the sound sensor to detect the working status of the speaker; the execution status of the light alarm is detected by the photoelectric sensor to detect the working status of the indicator light; the execution status of the screen display alarm is detected by the display driver module to detect the working status of the display screen; the execution status of the vibration alarm is detected by the vibration sensor to detect the working status of the vibration module; the monitoring results are fed back to the control system in real time;

[0124] S323: When an abnormality is detected in the execution of a certain alarm mode, the system will automatically switch to the backup alarm mode. Abnormalities include situations such as alarm device malfunction preventing alarm execution or alarm effects not meeting expectations. When an abnormality is detected in the execution of a certain alarm mode, the control system will automatically switch to the next higher priority alarm mode according to the pre-set backup plan. During the switching process, the fault information of the abnormal alarm mode will be recorded for subsequent maintenance. The switching of backup alarm modes follows a pre-set priority order to ensure that alarm information can be transmitted to operators in a timely and effective manner.

[0125] The beneficial effects of the above technical solution are as follows: by analyzing historical alarm records and current operating status, the most suitable alarm scheme is automatically adjusted, improving the effectiveness of alarms. Specifically, the analysis method can involve statistically analyzing the effectiveness data of various alarm methods under different environmental conditions to identify the most effective combination of alarm methods and parameter settings under specific conditions, further enhancing the reliability of POS machine anomaly alarms and improving user experience.

[0126] In another embodiment, such as Figure 3 As shown, a cash register malfunction alarm system includes:

[0127] The data monitoring module is used to acquire the operating data of the cash register when it is performing tasks, determine whether there are any abnormalities in the operating data, and generate abnormality detection data if there are any abnormalities.

[0128] The intelligent analysis module is used to analyze anomaly detection data based on a preset intelligent recognition model, determine the anomaly type, and identify the corresponding alarm scheme.

[0129] The alarm execution module is used to dynamically trigger a multi-mode collaborative alarm mechanism based on the corresponding alarm scheme and provide intelligent processing suggestions.

[0130] The working principle of the above technical solution is as follows: The data monitoring module includes:

[0131] The data acquisition submodule is used to collect real-time operating data of the cash register, including scanner status, payment system status, and network connection status.

[0132] The anomaly detection submodule is used to compare the collected operational data with preset normal operating thresholds to identify abnormal data points, and further includes:

[0133] The model building unit is used to build time series models of normal operating data;

[0134] The deviation calculation unit is used to calculate the deviation value between the current running data and the time series model;

[0135] An anomaly marking unit is used to mark data points as abnormal when the deviation value exceeds a preset threshold;

[0136] The anomaly data generation submodule is used to generate anomaly detection data based on the distribution characteristics of anomaly data points.

[0137] The intelligent analysis module includes:

[0138] The feature extraction submodule is used to input anomaly detection data into the intelligent recognition model and extract anomaly feature vectors;

[0139] The type matching submodule is used to determine the anomaly type by matching against a pre-defined anomaly type library based on anomaly feature vectors, and further includes:

[0140] Feature space construction unit, used to construct a multidimensional feature space for anomalous feature vectors;

[0141] The similarity calculation unit is used to calculate the similarity between the anomaly feature vector and the feature vectors of each anomaly type;

[0142] The type selection unit is used to select the anomaly type with the highest similarity as the final judgment result;

[0143] The scheme selection submodule is used to select the corresponding alarm scheme from the preset alarm scheme library based on the anomaly type and the current POS environment.

[0144] The alarm execution module includes:

[0145] The alarm configuration submodule is used to determine the combination of alarm triggering conditions and alarm methods according to the corresponding alarm scheme;

[0146] The alarm triggering submodule is used to monitor alarm triggering conditions in real time. When the conditions are met, multiple alarm methods are triggered simultaneously, including:

[0147] The sequence determination unit is used to determine the triggering order of alarm methods according to the alarm scheme. Alarm methods include sound, light, screen display and vibration.

[0148] The status monitoring unit is used to monitor the execution status of each alarm mode in real time;

[0149] The switching unit is used to automatically switch to the backup alarm mode when an abnormality is detected in the execution of a certain alarm mode;

[0150] It is recommended to generate a submodule to generate intelligent handling suggestions based on the exception type and the current checkout status and push them to the cashier's interface.

[0151] The beneficial effects of the above technical solution are as follows: The system is divided into a data monitoring module, an intelligent analysis module, and an alarm execution module, making the system structure clearer and facilitating development, maintenance, and upgrades. Modular design allows each functional module to work relatively independently, improving system stability and reliability. The collaborative work between modules forms a complete closed loop from data monitoring to anomaly analysis and alarm execution, ensuring the efficient execution of the anomaly alarm process.

[0152] In another embodiment, a cash register includes:

[0153] An anomaly alarm system is used to execute alarm methods for anomalies in the cash register.

[0154] The main control unit is connected to the anomaly alarm system and is used to perform cashier operations and provide operational status data to the anomaly alarm system.

[0155] The barcode scanner is used to scan product barcodes and transmit the data to the main control unit. The abnormal alarm system monitors whether the barcode scanner is scanning abnormally.

[0156] The touch screen is used to display cashier information and receive user input, and switches to a specific alarm interface when an alarm signal is received from the abnormal alarm system;

[0157] A multi-mode alarm device, including a sound alarm, indicator lights, and a vibration motor, is used to achieve multi-mode coordinated alarm based on the control of an abnormal alarm system.

[0158] The working principle of the above technical solution is as follows: The user (cashier) uses a barcode scanner to scan the barcode on the product. The barcode scanner transmits the scanned barcode data to the main control unit. After receiving the barcode data, the main control unit processes and updates the current payment information (such as product name, price, total amount, etc.), and displays this information in real time on the touch screen. The user can view the payment information, enter other necessary data (such as quantity, discount, etc.), or confirm the operation (such as completing the transaction) through the touch screen.

[0159] The anomaly alarm system continuously monitors the barcode scanner's scanning status to determine if any scanning anomalies exist (such as scan failure, unreadable barcode, continuous scan errors, etc.). Once an anomaly is detected, the alarm system immediately sends an anomaly signal to the main control unit. Upon receiving the anomaly signal, the main control unit triggers the alarm mechanism. The touchscreen display, upon receiving the alarm signal, automatically switches to a specific alarm interface, displaying anomaly information (such as "Scan anomaly, please check the barcode") and handling suggestions (such as "Please rescan or manually enter the barcode").

[0160] Simultaneously, the multi-mode alarm device activates coordinated alarms according to the control of the abnormal alarm system: Audible alarm: emits an alarm sound to alert the user to abnormal situations. Indicator light: flashes or displays a specific color (such as red) to visually indicate the abnormal state. Vibration motor: generates vibration to further enhance the alarm effect, ensuring that the user can perceive the alarm even in noisy environments.

[0161] Users become aware of the anomaly through the alarm interface on the touch screen and the prompts from the multi-mode alarm device, and can take appropriate measures based on the suggestions on the interface (such as rescanning the barcode, checking whether the barcode is damaged, etc.).

[0162] Users should take steps to resolve scanning anomalies based on the prompts on the alarm interface (such as changing the scanning angle, cleaning the barcode, or switching scanning modes). Once the anomaly is resolved, the alarm system detects that the barcode scanner has returned to normal scanning mode and sends a normal operation signal to the main control unit.

[0163] Upon receiving the recovery signal, the main control unit disables the alarm mechanism. The touchscreen display automatically switches back to the normal checkout interface, resuming the display of checkout information.

[0164] The multi-mode alarm device stops all alarm actions: the audible alarm stops sounding; indicator lights return to normal (e.g., off or green); and the vibration motor stops vibrating.

[0165] Users can continue with normal checkout operations, and the system will return to its state before the anomaly occurred.

[0166] The beneficial effects of the above technical solution are as follows: It organically integrates the anomaly alarm system with the main control unit of the cash register, barcode scanner, and other hardware components, constructing a comprehensive system that integrates cash register and anomaly alarm functions. The multi-mode alarm device includes an audible alarm, indicator lights, and a vibration motor, which can alert cashiers through multiple sensory channels, ensuring that anomalies are not ignored. The touchscreen display can switch to a specific alarm interface upon receiving an alarm signal, intuitively displaying anomaly information and handling suggestions to the cashier, improving the efficiency of anomaly handling.

[0167] In another embodiment, the cash register further includes:

[0168] Cash drawers are used to store cash and are connected to an anomaly alarm system to monitor for abnormal opening.

[0169] A printer used to print cash register receipts and connected to an anomaly alarm system to monitor printing anomalies.

[0170] The working principle of the above technical solution is as follows: Transaction processing is the core function of a cash register, involving the input of product information, the calculation and display of payment amounts, the selection of payment methods, and the printing of transaction receipts. Cash registers are typically equipped with two displays: a main display facing the cashier and a secondary display facing the customer, and the content displayed on both is adjusted according to the viewing angle.

[0171] The cashier enters product information via keyboard on the main display screen or scans product barcodes using a barcode scanner. The system automatically calculates the total amount and displays detailed transaction information (such as product name, unit price, quantity, and total amount) on the main display screen. Simultaneously, a secondary display screen faces the customer, showing simplified payment prompts such as "Please pay XX yuan" to inform the customer of the amount due.

[0172] Customers select their payment method (such as cash, credit card, or mobile payment) based on the amount displayed on the secondary screen. The cashier then selects the corresponding payment option on the main screen.

[0173] If the customer chooses to pay in cash: The cashier presses the "Cash Payment" button on the main display screen, and the system automatically opens the cash drawer. The main display screen shows "Please pay XX yuan," prompting the cashier to collect the customer's cash. The customer hands the cash to the cashier, who places the cash into the cash drawer and presses the "Confirm" button. The system calculates the change, and the main display screen shows "Change XX yuan," while the secondary display screen updates accordingly. The cashier then retrieves the corresponding amount of change from the cash drawer and gives it to the customer.

[0174] After payment is completed, the system confirms the transaction is finished, and the printer automatically prints a receipt. The receipt contains detailed transaction information, such as the item list, total amount, payment method, and change. The main display shows "Transaction Completed," and the secondary display shows "Thank you for your patronage" or a similar message. The cash drawer is an important component of the cash register, used to store cash and connected to an alarm system to ensure cash security.

[0175] In cash transactions, the cashier opens the cash drawer by pressing the "Cash Payment" button, places the customer's cash into a designated compartment, and retrieves change. The cash drawer is typically divided into multiple compartments for storing banknotes and coins of different denominations. Each time the cash drawer is opened, the system records the opening time and associated transaction information (such as transaction number and amount), storing the data in the cash register's log. If the cash drawer is opened outside of transaction hours or without authorization (e.g., by physical prying or abnormal operation), the anomaly alarm system detects this and immediately triggers an alarm, notifying the administrator or security system to prevent cash theft. The anomaly alarm system connects to the cash drawer and printer, monitoring device status in real time to ensure the security and reliability of the transaction process. The system continuously monitors the cash drawer's opening status. If abnormal opening is detected (e.g., opening the cash drawer without following the normal transaction process), the system triggers an audible alarm or sends an alarm notification to the management terminal. The system checks the printer's operational status while printing receipts. If an abnormal situation occurs, such as running out of paper, paper jam, or hardware failure, the abnormal alarm system will be triggered, and the main display screen may show a message such as "Printer malfunction, please check" to remind the cashier to handle it in time (such as replacing paper or repairing the equipment).

[0176] The beneficial effects of the above technical solution are as follows: Adding a cash drawer monitoring function enables timely detection of abnormal cash drawer opening, preventing cash loss or theft and improving the security of the cash register. Monitoring the printer enables timely detection of printing anomalies, ensuring the normal printing of receipts and improving the integrity and reliability of the checkout process. Connecting the cash drawer and printer to the anomaly alarm system expands the scope of anomaly monitoring, making the cash register's anomaly alarm function more comprehensive and complete.

[0177] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from the spirit and scope of this invention.

Claims

1. A method for alarming abnormalities in a cash register, characterized in that, include: S1: Obtain the operating data of the cash register when it is performing a task, determine whether there are any abnormalities in the operating data, and generate abnormality detection data if there are any abnormalities. S2: Based on the preset intelligent recognition model, analyze the anomaly detection data, determine the anomaly type, and determine the corresponding alarm scheme; S3: Based on the corresponding alarm scheme, dynamically trigger a multi-mode collaborative alarm mechanism and provide intelligent processing suggestions.

2. The cash register malfunction alarm method according to claim 1, characterized in that, Step S1 includes: S11: Real-time collection of cash register operation data, including scanner status, payment system status, and network connection status; S12: Compare the collected operational data with the preset normal operation threshold to identify abnormal data points; S13: Generate anomaly detection data based on the distribution characteristics of abnormal data points.

3. The cash register abnormality alarm method according to claim 1, characterized in that, Step S2 includes: S21: Input the anomaly detection data into the intelligent recognition model and extract the anomaly feature vector; S22: Based on the abnormal feature vector, match it in the preset abnormal type library to determine the abnormal type; S23: Select the corresponding alarm scheme from the preset alarm scheme library based on the anomaly type and the current POS environment.

4. The cash register malfunction alarm method according to claim 1, characterized in that, Step S3 includes: S31: Determine the combination of alarm triggering conditions and alarm methods according to the corresponding alarm scheme; S32: Real-time monitoring of alarm triggering conditions; when the conditions are met, multiple alarm modes are triggered simultaneously. S33: Based on the exception type and the current checkout status, generate intelligent handling suggestions and push them to the cashier's interface.

5. The cash register malfunction alarm method according to claim 2, characterized in that, Step S12 includes: S121: Establish a time series model of normal operation data; S122: Calculate the deviation between the current running data and the time series model; S123: When the deviation value exceeds the preset threshold, it is marked as an abnormal data point.

6. The cash register abnormality alarm method according to claim 3, characterized in that, Step S22 includes: S221: Construct a multidimensional feature space for abnormal feature vectors; S222: Calculate the similarity between the anomaly feature vector and the feature vectors of each anomaly type; S223: Select the anomaly type with the highest similarity as the final judgment result.

7. The cash register malfunction alarm method according to claim 4, characterized in that, Step S32 includes: S321: Determine the triggering sequence of alarm methods according to the alarm scheme. Alarm methods include sound, light, screen display and vibration. S322: Real-time monitoring of the execution status of each alarm mode; S323: When an abnormality is detected in the execution of a certain alarm mode, the system will automatically switch to the backup alarm mode.

8. A cash register malfunction alarm system employing the cash register malfunction alarm method as described in any one of claims 1-7, characterized in that, include: The data monitoring module is used to acquire the operating data of the cash register when it is performing tasks, determine whether there are any abnormalities in the operating data, and generate abnormality detection data if there are any abnormalities. The intelligent analysis module is used to analyze anomaly detection data based on a preset intelligent recognition model, determine the anomaly type, and identify the corresponding alarm scheme. The alarm execution module is used to dynamically trigger a multi-mode collaborative alarm mechanism based on the corresponding alarm scheme and provide intelligent processing suggestions.

9. A cash register employing the cash register malfunction alarm system as described in claim 8, characterized in that, include: The main control unit is connected to the anomaly alarm system and is used to perform cashier operations and provide operational status data to the anomaly alarm system. The barcode scanner is used to scan product barcodes and transmit the data to the main control unit. The abnormal alarm system monitors whether the barcode scanner is scanning abnormally. The touch screen is used to display cashier information and receive user input, and switches to a specific alarm interface when an alarm signal is received from the abnormal alarm system; A multi-mode alarm device, including a sound alarm, indicator lights, and a vibration motor, is used to achieve multi-mode coordinated alarm based on the control of an abnormal alarm system.

10. The cash register according to claim 9, characterized in that, Also includes: Cash drawers are used to store cash and are connected to an anomaly alarm system to monitor for abnormal opening. A printer used to print cash register receipts and connected to an anomaly alarm system to monitor printing anomalies.

Images