KR102443470B1 - Product classification method and system based on product volume value and corrected weight value - Google Patents

Abstract

The present invention relates to a method and system for classifying parcel delivery products, specifically, measuring the volume of goods moved by a conveyor, and separately weighing and weighing goods that require weight measurement according to the measured volume value It relates to a product classification method and system for allowing each product to be classified into a correct suit by performing weight correction.

Description

Product classification method and system based on the volume value of the product and the corrected weight value

The present invention relates to a method and system for classifying parcel delivery products, specifically, measuring the volume of goods moved by a conveyor, and separately weighing and weighing goods that require weight measurement according to the measured volume value It relates to a product classification method and system for allowing each product to be classified into a correct suit by performing weight correction.

The expansion of the e-commerce infrastructure and the full-fledged non-face-to-face socialization are driving the growth of the parcel delivery market, and this trend is expected to continue in the future. In line with this trend, many courier companies are interested in a product sorting system composed of automated equipment, and the number of workplaces that have installed an actual product sorting system is increasing. However, the product classification systems commercialized so far are not yet fully automated, and in particular, the loading and unloading of parcels is mostly done manually by humans.

On the other hand, unavoidable human errors inevitably occur during the loading and unloading operations that are manually performed by humans, and since such human errors can sometimes lead to fatal accidents in the workplace, the need for prevention is urgent. As a representative example, there may be cases in which products with very heavy weight (eg dumbbells, etc.) are classified as small product lines even though the volume (volume) is small because the weight of the parcel delivery product is not recognized in advance during the disembarkation operation. In this case, a heavy product may damage the equipment in the small product line section, or the product may fall on the worker and lead to a serious accident, such as a serious injury to the worker.

The present invention has been proposed in view of such a problem, and relates to a product classification method, and a system for the same, by measuring and correcting not only the volume of the product but also the weight if necessary, so that the product can be classified into a correct suit. In addition, the present invention was invented to provide additional technical elements that cannot be easily invented by a person skilled in the art in addition to solving the above technical problems.

Republic of Korea Patent No. 10-0861382 (Registered on Sep. 25, 2008)

The present invention not only measures the volume value of the parcel delivery product, but also measures and corrects the weight to obtain an accurate weight value so that the parcel delivery product can be classified into the correct suit.

In particular, an object of the present invention is to select products having a heavy weight even though they have a small volume and classify them as products of a handling care, thereby preventing accidents in the workplace.

In addition, an object of the present invention is to enable the weight of the product to be measured relatively accurately by providing a plurality of load cells, and furthermore, to accurately obtain the weight of the product by enabling weight correction based on the measured weight values.

In addition, the present invention is characterized in that the correct product weight can be grasped by allowing the product weight correction to be made by an artificial intelligence algorithm.

On the other hand, the technical problems of the present invention are not limited to the technical problems mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art from the following description.

The present invention is to solve the above problems, the product classification method according to the present invention comprises the steps of (a) measuring the volume of the product; (b) determining whether the product is a product requiring measurement of weight; (c) measuring the weight of the product when it is determined as a product requiring measurement of weight; (d) classifying the product as a predetermined suit if the weight of the product is greater than or equal to a preset value as a result of the measurement.

In addition, in the product classification method, the step (c) comprises the steps of measuring the weight of the product; and correcting the weight of the product; may include.

In addition, before the step of measuring the weight of the product in the product classification method, the step of detecting that the product enters the weight measurement unit for weight measurement; may further include.

In addition, the step of measuring the weight of the product in the product classification method includes: (c21) acquiring a plurality of time series weight values; (c22) defining n weight values as a first window; (c23) defining a maximum value within the first window as a first maximum value; (c24) defining n weight values as the i-th window starting with the next-order weight value of the first weight value in the i-1th window (i is an integer that increases according to the number of regressions starting from 2); (c25) defining a maximum value within the ith window as an ith maximum value; (c26) When the amount of change between the i-1 maximum value and the i-th maximum value is a value of 0 or less, the i-1 maximum value is defined as a peak value, and when the amount of change is a value greater than 0 (c24) It may include; returning to the step.

In addition, the step of measuring the weight of the product in the product classification method may be characterized in that the steps (c24) to (c26) are repeated until the peak value is defined.

In addition, the step of correcting the weight of the product in the product classification method comprises: (c31) inputting the peak value and m weight values before the peak value (m is an integer greater than or equal to 2) into the artificial intelligence algorithm. step; (c32) performing weight value correction using the artificial intelligence algorithm; may include.

In addition, in the product classification method, the step (b) may be a step of determining whether the product is a small or medium-sized product according to whether the volume of the product is included in a preset range.

In addition, in the step (d) of the product classification method, if the weight of the product is greater than or equal to a preset value, it may be characterized in that the product is classified into a suit for classifying different types of products.

On the other hand, in the method of measuring the weight of a product passing through a load cell according to another embodiment of the present invention, (s1) n weight values are k-th (k is 2, starting from 2, depending on the number of regressions) an increasing integer) defining a window; (s2) defining a maximum value within the k-th window as a k-th maximum value; (s3) As a result of comparing the k-th maximum value with the k-1 th maximum value, which is the maximum value of the k-1 th window, if the change amount is 0 or less, the k-th maximum value is defined as a peak value, and the change amount If this value is greater than 0, the method includes returning to the step (s1), wherein at least one of the n weight values in the step (s1) after the regression is time-series compared to the weight value at the first time in the k-th window It may be characterized as a subordinate.

In addition, the method of measuring the product weight of the product passing through the load cell may be characterized in that the steps (s1) to (s3) are repeated until the peak value is defined.

Also, at this time, after the peak value is defined in step (s3), the peak value and m weight values before the peak value (m is an integer of 2 or more) are input to the artificial intelligence algorithm to correct the weight value. performing; may further include.

On the other hand, the product classification system according to another embodiment of the present invention, a volume measurement unit for measuring the volume of the products passing through the transport means; a weight measuring unit for measuring the weight of a product classified as a product requiring weight measurement based on a measured volume value among products passing through the transport means; It may include; a sorting unit for classifying the product into any one suit among a plurality of suits according to the data received from the volume measurement unit or the weight measurement unit.

In addition, in the product classification system, the weight measurement unit may measure the weight of a product whose volume is included in a preset volume range of small and medium-sized products.

In addition, in the product classification system, the sorting unit may control the product to be classified into a predetermined suit when the weight of the product measured by the weight measurement unit is equal to or greater than a preset value.

The product classification system may further include a sensing unit that is provided prior to the weight measurement unit and detects product entry into the weight measurement unit.

In addition, in the product classification system, the weight measurement unit may include: (t21) receiving a plurality of time series weight values measured by the weight measurement unit; (t22) defining n weight values (n is an integer greater than or equal to 2) as a first window; (t23) defining a maximum value within the first window as a first maximum value; (t24) defining n weight values as the i-th window starting with the next-order weight value of the first weight value in the i-1th window (where i is an integer that increases according to the number of regressions starting from 2); (t25) defining a maximum value within the ith window as an ith maximum value; (t26) When the amount of change between the i-1 maximum value and the i-th maximum value is 0 or less, the i-1 maximum value is defined as a peak value, and when the amount of change is a value greater than 0 (t24) It is characterized in that the peak value is defined by performing the step of regressing, wherein at least one of the n weight values in the step (t24) after the regression is time-series compared to the weight value of the first time in the k-th window It may be characterized as a subordinate.

According to the present invention, since it is possible to classify the product by measuring the volume and, if necessary, the weight of the product being transported on the conveyor, it is possible to accurately classify the product.

In addition, according to the present invention, since it is possible to classify goods that are small in volume and heavy in weight, there is an effect that can prevent accidents in the workplace that may occur due to these goods in advance.

In addition, according to the present invention, there is an effect that more accurate classification is possible by correcting the measured weight using an artificial intelligence algorithm.

In addition, according to the present invention, there is an effect of increasing the overall speed and efficiency of the sorting operation because workers who put the goods do not have to check the weight of the goods and then sort them.

On the other hand, the effects of the present invention are not limited to those mentioned above, and other technical effects not mentioned will be clearly understood by those skilled in the art from the following description.

1 is a schematic diagram of a product classification system implemented in a conventional unloading workshop, and FIG. 2 is a diagram illustrating a conventional product classification method.
3 is a diagram illustrating a product classification system according to the present invention, and FIG. 4 is a diagram illustrating a product classification process according to the system.
5 shows detailed steps of the weight measurement step S103 of FIG. 4 , and shows steps for measuring and correcting the weight of the product.
6 is a view showing the subdivision of the weight measurement step of the product, Figure 7 is a subdivided view showing the weight correction step.

The purpose and technical configuration of the present invention, and details regarding the operational effects thereof will be more clearly understood by the following detailed description based on the accompanying drawings in the specification of the present invention. An embodiment according to the present invention will be described in detail with reference to the accompanying drawings.

The embodiments disclosed herein should not be construed or used as limiting the scope of the present invention. It is natural for those skilled in the art that the description including the embodiments of the present specification will have various applications. Accordingly, any embodiments described in the detailed description of the present invention are illustrative for better describing the present invention and are not intended to limit the scope of the present invention to the embodiments.

The functional blocks shown in the drawings and described below are merely examples of possible implementations. Other functional blocks may be used in other implementations without departing from the spirit and scope of the detailed description. Further, although one or more functional blocks of the present invention are represented as separate blocks, one or more of the functional blocks of the present invention may be combinations of various hardware and software configurations that perform the same function.

In addition, the expression that includes certain components is an expression of “open type” and merely refers to the existence of the corresponding components, and should not be construed as excluding additional components.

Furthermore, when it is said that a component is “connected” or “connected” to another component, it may be directly connected or connected to the other component, but it should be understood that other components may exist in between. do.

1 is a view showing a conventional product classification state, in which workers who put products from a container to a conveyor, and a product classification system for classifying the put products are shown.

The conventional product classification system is largely composed of a volume measuring unit 11 and a sorting unit 15, and in addition, a control 19 unit for controlling the volume measuring unit 11 and the sorting unit 15 is further included. can

2 shows a process of product classification in a conventional product classification system. Referring to this, in the related art, the volume of a product transferred through a conveyor is simply measured (S11), and as a result of the measurement, whether the product is a large product, small or medium-sized product It is determined whether the product is recognized or a different type product (S12), and the sorting unit 15 is driven (S13) according to the determination result to transfer the product to each chute, thereby classifying the product.

However, in such a conventional product classification system and method, since only the volume is acquired as information necessary for product classification, product classification that leads to a large accident has been sometimes made, especially when a product with a small volume has a heavy weight. there has been

In response to this, the product classification system according to the present invention was further provided with a means for measuring the weight of the product to solve the above problems.

3 shows a product classification system according to the present invention, according to which the product classification system may include a volume measurement unit 110 , a weight measurement unit 130 , a sorting unit 150 , and a control unit 170 . and may further include a product detection unit 190 .

The volume measuring unit 110 may be understood as a means for measuring the volume of the product when the product is transported and passed through the conveyor. Volume refers to the size that an arbitrary object occupies in space, and the measurement of the volume can be made by arbitrary calculation from the numerical values for the product shape obtainable by various detection sensors, or an image of a passing product It can be done in various ways, such as can be done according to the analysis.

On the other hand, the volume measuring unit 110 measures the volume of the product, and then, according to the measured volume value, is implemented to determine whether the corresponding product is a product that requires weight measurement or a product that does not require weight measurement. Alternatively, unlike the above, the volume measurement unit 110 transmits the measured volume value to the control unit 170 to be described later, thereby allowing the control unit 170 to determine whether the product is a product requiring weight measurement. have. For reference, the volume measurement unit 110 or the control unit 170 may internally store reference values for the classification criteria of the product, for example, when the measured volume value is greater than x, the product is a variant product, x to y Reference values may be stored to classify a large product when it is within the range of y and a small and medium product when it is smaller than y.

Meanwhile, the volume measuring unit 110 may include a function for not only measuring the volume of a non-product but also acquiring other product information(s) for the product when an arbitrary product enters the product. For example, the volume measuring unit 110 may further include means for acquiring arbitrary information about a passing product, such as an RFID reader, a QR reader, a barcode scanner, a camera, etc. Information data that can be referenced in the classification process can be generated.

The weight measurement unit 130 is a means for measuring the weight of the goods conveyed and passed through the conveyor. The weight refers to the degree of weight of an object, and the measurement of the weight may be performed by a weight sensing means such as a load cell on a path through which the product is transported. For reference, in FIG. 3 , the load cells are shown as four small circles in the weight measuring unit 130 . A load cell is a device for measuring weight strictly, and the device undergoes deformation such as being compressed or stretched when subjected to pressure by weight. It is a means that made weight measurement possible by changing it. On the other hand, in the present invention, the weight measuring unit 130 can measure the weight of a product even if in principle only one load cell is provided, but preferably, it may include at least four load cells for more accurate measurement.

On the other hand, the weight measurement unit 130 does not measure all the weights of all the products transferred through the conveyor, and may be implemented to measure only the weight of the products that need to be measured, especially among the products conveyed through the conveyor. Although it will be described later, the weight measurement unit 130 does not measure the weight of products having a volume corresponding to a large or deformed product as a result of the measurement in the previous volume measurement unit 110, It may be implemented to measure the weight only for products having a volume. In this case, the standard for the different, large, and small and medium-sized volumes may of course be predefined in advance.

In addition, the weight measurement unit 130 may transmit the measured weight value to the control unit 170, which will be described later, so that the control unit 170 may directly perform a weight correction operation. Alternatively, the weight measurement unit 130 may have its own calculation means and may be implemented to perform correction on a weight value measured for an arbitrary product.

Next, the sorting unit 150 classifies the goods transported through the conveyor into appropriate suits according to the control command of the control unit 170 or according to the data received from the volume measuring unit 110 or the weight measuring unit 130 . is a configuration that The sorting unit 150, preferably, may be implemented in the form of a wheel sorter in which a plurality of direction changeable rollers (direction changing rollers) are disposed, and by controlling the rotation angle of the direction changing roller, each product is divided into a plurality of It can be driven to be classified into an appropriate suit among the suits. However, it is only one embodiment that the sorting unit 150 is implemented including a wheel sorter, and if the sorting unit 150 can sort goods moving through a conveyor, in addition to the wheel sorter, a cross belt sorter, a tray Any other form of implementation, including the like, would be irrelevant.

Next, the control unit 170 may be understood as a configuration that generally controls the aforementioned volume measurement unit 110 , weight measurement unit 130 , and sorting unit 150 . The control unit 170 referred to in this detailed description may be understood as a configuration capable of controlling the volume measurement unit 110 , the weight measurement unit 130 , and the sorting unit 150 as a whole, but within the actually implemented system It is understood that the volume measurement unit 110 , the weight measurement unit 130 , and the sorting unit 150 may be driven by directly exchanging data with each other while the configuration of the control unit 170 is omitted. The control unit 170 may be implemented as a single hardware device including a central processing unit (CPU) and a memory in an embodiment. In this case, the central processing unit may also be called a controller, a microcontroller, a microprocessor, a microcomputer, or the like. In addition, the central processing unit may be implemented by hardware, firmware, software, or a combination thereof. When implemented using hardware, an ASIC (application specific integrated circuit) or DSP (digital signal processor) , DSPD (digital signal processing device), PLD (programmable logic device), FPGA (field programmable gate array), etc., when implemented using firmware or software, a module, procedure or function that performs the above functions or operations, etc. Firmware or software may be configured to include In addition, the memory includes ROM (Read Only Memory), RAM (Random Access Memory), EPROM (Erasable Programmable Read Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), flash memory, SRAM (Static RAM), It may be implemented as a hard disk drive (HDD), a solid state drive (SSD), or the like.

Finally, the product classification system according to the present invention may further include a product detection unit 190 provided before the weight measurement unit 130 . The product detection unit 190 is configured to detect that the product input through the conveyor enters the weight measurement unit 130, and is installed at any point in the movement path of the product or installed on the side of the conveyor as shown in the drawing. It may be provided to detect the moment the product passes by. The product detection unit 190 may preferably be implemented as an optical sensor, but as long as it can detect the entry of the product, there is no limitation on the means of implementation. On the other hand, in another embodiment, the product detection unit 190 may be implemented to acquire brief information about the entering product as needed, not necessarily only detecting whether the product has entered. For example, the product detection unit 190 reads a series of codes attached to the product or measures the length and size of the product based on the time between the time when the product starts to pass and the time when the product passes through. It can also be implemented so that it is possible to obtain brief information about it.

On the other hand, when the entry of the product is detected by the product detection unit 190, the product detection unit 190 directly transmits a detection signal to the weight measurement unit 130 to cause the weight measurement unit 130 to enter the product. Alternatively, the product detection unit 190 may be implemented to transmit a detection signal to the control unit 170 so that the control unit 170 drives and controls the weight measurement unit 130 according to the product entry timing. have. As such, the product detection unit 190 is provided to determine the entry point of the product and transmit the signal to the weight measurement unit 130 or the control unit 170 so that the weight measurement of the product being transported can be performed at an accurate timing. can be

The overall outline of the product classification system according to the present invention has been looked at with reference to FIG. 3 above.

4 is a view showing the product classification methods according to the present invention in order. Referring to FIG. 4 , the product classification method may start from the step ( S101 ) of measuring the volume of the first transported product. This step is executed when the product transported through the conveyor passes through the volume measuring unit 110, and the volume measuring unit 110 includes various types of sensors or photographing means (camera, etc.) provided therein. can be used to measure how much space a product takes up, that is, how much volume it has.

After step S101, a step of determining whether the product is a product requiring weight measurement based on the measured volume value may be executed. This step is a step that can be performed by the volume measuring unit 110 or the control unit 170 receiving the volume value from the volume measuring unit 110. For example, in this step, based on the measured volume value of the product, the It can be determined whether the product is a small and medium-sized product, a large product, or a variant product. If the product is judged to be a different type or a large product in terms of volume value, there is no need to measure the weight of the product, and you can proceed with a different type or a large suit, respectively. It is necessary to measure the weight because it is necessary to separately sort products that have a small volume but have a lot of weight. 4 shows the step of determining whether the product measured by volume in step S102 is a different type, large size, or small or medium size, but this step is actually a product requiring weight measurement based on the volume value for the product It can be understood as a step of determining whether or not it is recognized.

In step S102, if it is determined that the product is a product that does not require measurement of weight (large or deformed product in the drawing), the product may be directly sorted by the sorting unit 150 while omitting the weight measurement step (S105).

However, if it is determined that the product needs to be weighed in step S102, the weighing step (S103) may be performed subsequently. This step is a step executed by the weight measuring unit 130, and as mentioned in the description in FIG. 3 above, the weight may be measured when the product passes over the load cell.

On the other hand, more detailed steps of step S103 are shown in FIG. 5. Referring to FIG. 5, step S103 is again a step of detecting entry into the product (S1031), a step of measuring the weight of the product (S1032), and the measured weight. It may include a step of correcting (S1033). The step (S1031) of detecting the product entry may be performed by the product detection unit 190 provided before the weight measurement unit 130, and the step (S1032) of measuring the weight of the product is the weight measurement unit 130. and correcting the measured weight ( S1033 ) may be performed by the weight measurement unit 130 or the control unit 170 . Of these, the step of measuring the weight (S1032) and the step of correcting the measured weight (S1033) will be described in more detail later with reference to FIGS. 6 and 7, and here, back to the description of FIG. 4 I would like to finish the explanation.

Referring back to FIG. 4 , after step S103 , a step S104 of determining whether the measured and corrected weight is greater than or equal to a preset value (M kg) may be performed. This step is a step that can be performed by the weight measurement unit 130 or the control unit 170, and this step is a product requiring special attention in handling, that is, handling if the measured and corrected weight value is greater than or equal to the preset value. This is a step to classify as a caution product. As an example, the step of determining whether the measured and corrected product weight is 20 kg or more may be performed, and if the determination result is less than 20 kg, it is determined as a small and medium-sized product, and if it is 20 kg or more, it is determined as a variant product, and in the sorting unit 150 Classification (S105) may be made according to the above result.

An embodiment of the product classification method according to the present invention has been described with reference to FIG. 4 above. As described above, in the present invention, not only the volume is measured to determine the degree of care in handling the product, but the volume measurement and weight measurement (correction) are performed to have a heavy weight even though the volume is small, which can lead to accidents in the workplace. It is an important feature to distinguish existing products.

Hereinafter, detailed descriptions of the previously delayed weight measurement step (S1032) and weight correction step (S1033) will be continued. For reference, the weight measurement step ( S1032 ) and the weight correction step ( S1033 ) may be performed by the weight measurement unit 130 or the control unit 170 , preferably implemented to be performed by the weight measurement unit 130 . can

First, Figure 6 lists detailed steps of the weight measurement step (S1032). Referring to FIG. 6 , the step of measuring the weight ( S1032 ) may start from the step of first acquiring a plurality of weight values ( S10321 ). When the product passes on the weight measuring unit 130, more precisely, when the product passes over the load cell, weight values can be obtained in a very short time unit, for example, the weight value is 20 ms (in 1 second). 50) units.

After step S10321, n (n is any positive number) weight values (preferably consecutive weight values) are defined as the first window (which may also be referred to as the first window or the first window). Step S10322 may be performed, and then step S10323 of defining a maximum value among the weight values of the first window as a first maximum value may be performed. At this time, the window can be understood as a set including a plurality of weight values, and as will be understood later, the definition of the window concept is to obtain a maximum value among a plurality of time-series continuous values.

On the other hand, after the first window and the first maximum value are defined, n of the weight values (in this case, the n weight values start with the next weight value in time series from the first weight value in the first window) The step of defining the weight value of the second window as the value) may be performed, and then the step of defining the maximum value among the weight values of the second window as the second maximum value may be performed.

On the other hand, after the second window and the second maximum value are defined, the amount of change of the first window with the first maximum value is calculated. value), and if the amount of change is greater than 0, define n weight values as the third window in a sliding method again, and define the maximum value among them as the third maximum value to compare the change amount with the second maximum value. will do If the amount of change between the maximum value of the previous step window and the maximum value of the current step window is greater than 0, the above calculation process is repeated. . That is, after the first window and the first maximum value are obtained, defining the n weight values as the i-th window (i is an integer greater than or equal to 2) (S10324), finding and defining the maximum value within the i-th window (S10324) S10325), determining whether the change amount is 0 or negative by comparing the i-1 maximum value with the i-th maximum value (S10326), if the change amount is 0 or negative, setting the i-1 maximum value as the peak value (S10327) and, if the amount of change is greater than 0, a step of returning to step S10324 and repeating the calculation may be performed. In this case, the peak value may be the final weight measurement value of the product before the correction is made.

As an example, when it is assumed that weight values a1 to a50 are obtained for an arbitrary product, 45 weight values from a1 to a45 may be included in the first window, of which a45 is to be defined as the first maximum value. In addition, 45 weight values from a2 to a46 may be included in the second window, and among them, a46 may be defined as the second maximum value. Also, the amount of change between the first maximum value a45 and the second maximum value a46 may be calculated, and if a46 has a larger value and the change amount is negative, a45 may be determined as a peak value, When the variation is greater than 0 due to a large value, the process of finding a third window including 45 weight values from a3 to a47 and a third maximum value that is a maximum value within the third window may be repeated.

Meanwhile, after the weight measurement method of FIG. 6 is performed, the measured weight value may be corrected, which is illustrated in FIG. 7 .

7 is a subdivision of the weight correction step ( S1033 ) of FIG. 5 , and when referring to this, the weight correction step is the first determined peak value and the weight of m (m is an arbitrary positive number) obtained immediately before the peak value on the time series. It may start from the step S10331 in which values are input to the artificial intelligence algorithm. For example, when the peak value is a45, 10 weight values ranging from a45 to a36 may be input to the artificial intelligence algorithm.

After the peak value and the m weight values immediately preceding it are input to the artificial intelligence algorithm, the weight value may be corrected by the artificial intelligence algorithm. (S10332) In this case, the AI algorithm may be an LSTM that shows good performance on time series data, and in order to generalize the performance of the AI algorithm model and increase the reliability of the result, the Stacked LSTM and Batch Normalization technique in which LSTMs are stacked in multiple layers will be used. can In summary, a plurality of weight values (the peak value and m immediately before it) may be input to an artificial intelligence algorithm (eg, LSTM) in time series order, and as an output, the previously input weight may be input. Based on the values, the predicted next-order weight value may be obtained in consideration of the trend of the current input (a time-series characteristic obtained by learning by an artificial intelligence algorithm (LSTM)).

Meanwhile, the weight value corrected in step S10332 may be finally determined as the weight value of the product passing through the weight measurement unit 130 ( S10333 ).

For reference, LSTM (Long Short-Term Memory) is a type of RNN and is executed with similar logic to RNN, but LSTM is different from RNN in that it has a function of preserving long-term memory. LSTM has the advantage that it is easier to refer to long-term memory compared to general DNN by controlling to input, store, and output data coming through the input node in units of time. There is an advantage that can be expected to improve the recognition rate.

As described above, a product classification method and a system for the same according to the present invention have been described. On the other hand, the present invention is not limited to the specific embodiments and applications described above, and various modifications are carried out by those skilled in the art to which the invention pertains without departing from the gist of the invention as claimed in the claims. Of course, these modifications are not to be understood as being distinct from the spirit or vision of the present invention.

In particular, the components implementing the technical features of the present invention included in the block diagrams and flowcharts shown in the accompanying drawings herein mean logical boundaries between the components. However, according to an embodiment of software or hardware, the illustrated components and their functions are executed in the form of stand-alone software modules, monolithic software structures, codes, services, and combinations thereof, and can execute stored program codes, instructions, etc. Since the functions may be implemented by being stored in an executable medium in a computer having a processor, all such embodiments should also be regarded as falling within the scope of the present invention.

Accordingly, although the accompanying drawings and descriptions thereof describe the technical features of the present invention, they should not be simply inferred unless a specific arrangement of software for implementing these technical features is clearly mentioned. That is, various embodiments described above may exist, and since such embodiments may be partially modified while retaining the same technical characteristics as those of the present invention, this should also be regarded as falling within the scope of the present invention.

In addition, in the case of a flowchart, the operations are depicted in the drawings in a specific order, but this is illustrated in order to obtain the most desirable result. should not be construed as being In certain cases, multitasking and parallel processing can be advantageous. In addition, the separation of various system components of the above-described embodiments should not be construed as requiring such separation in all embodiments, and the program components and systems described are generally integrated together into a single software product or in multiple software products. It should be understood that it may be packaged.

As such, this specification is not intended to limit the invention by the specific terminology presented. Therefore, although the present invention has been described in detail with reference to the above-described embodiments, those of ordinary skill in the art to which the present invention pertains may modify, change and deformation can be applied. The scope of the present invention is indicated by the following claims rather than the above detailed description, and all changes or modifications derived from the meaning and scope of the claims and their equivalent concepts are interpreted as being included in the scope of the present invention. should be

11 (conventional) volume measurement unit
15 (conventional) sorting unit
110 Volume measuring unit 130 Weight measuring unit 150 Sorting unit
170 Control unit 190 Product detection unit

Claims (16)

In the method of classifying a product,
(a) measuring the volume of the product;
(b) determining whether the product is a product requiring weight measurement;
(c) when it is determined that the product requires weight measurement, measuring the weight of the product;
(d) classifying the product into a predetermined suit if the weight of the product is greater than or equal to a preset value as a result of the measurement;
containing,
How to classify goods.
According to claim 1,
Step (c) is,
measuring the weight of the product; and
correcting the weight of the product;
characterized in that it comprises,
How to classify goods.
3. The method of claim 2,
Before the step of measuring the weight of the product,
detecting that the product enters the weight measurement unit for weight measurement;
characterized in that it further comprises,
How to classify goods.
3. The method of claim 2,
The step of measuring the weight of the product,
(c21) acquiring a plurality of time series weight values;
(c22) defining n weight values as a first window;
(c23) defining a maximum value within the first window as a first maximum value;
(c24) defining n weight values as the i-th window starting with the next-order weight value of the first weight value in the i-1th window (i is an integer that increases according to the number of regressions starting from 2);
(c25) defining a maximum value within the ith window as an ith maximum value;
(c26) When the amount of change between the i-1 maximum value and the i-th maximum value is a value of 0 or less, the i-1 maximum value is defined as a peak value, and when the amount of change is a value greater than 0 (c24) returning to the stage;
characterized in that it comprises,
How to classify goods.
5. The method of claim 4,
The step of measuring the weight of the product,
characterized in that the steps (c24) to (c26) are repeated until the peak value is defined,
How to classify goods.
6. The method of claim 5,
The step of correcting the weight of the product,
(c31) inputting the peak value and m weight values before the peak value (m is an integer greater than or equal to 2) into an artificial intelligence algorithm;
(c32) performing weight value correction using the artificial intelligence algorithm;
characterized in that it comprises,
How to classify goods.
According to claim 1,
Step (b) is,
characterized in that the step of determining whether the product is a small or medium-sized product according to whether the volume of the product is included in a preset range,
How to classify goods.
. According to claim 1,
Step (d) is,
If the weight of the product is greater than or equal to a preset value, characterized in that it is classified as a suit for classifying different types of products,
How to classify goods.
In the method of measuring the product weight of the product passing through a load cell,
(s1) defining n weight values as a k-th window (k is an integer starting from 2 and increasing according to the number of regressions);
(s2) defining a maximum value within the k-th window as a k-th maximum value;
(s3) As a result of comparing the k-th maximum value with the k-1 th maximum value, which is the maximum value of the k-1 th window, if the change amount is 0 or less, the k-th maximum value is defined as a peak value, and the change amount If this value is greater than 0, including the step of returning to step (s1),
At least one of the n weight values in the step (s1) after regression is characterized in that it has a lower priority in time series than the weight value of the first time in the k-th window,
How to weigh goods.
10. The method of claim 9,
characterized in that the steps (s1) to (s3) are repeated until the peak value is defined,
How to weigh goods.
11. The method of claim 10,
After the peak value is defined in step (s3),
performing weight value correction by inputting the peak value and m weight values before the peak value (m is an integer greater than or equal to 2) into an artificial intelligence algorithm;
further comprising,
How to weigh goods.
In the product classification system,
a volume measuring unit for measuring the volume of goods passing through the conveying means;
a weight measuring unit for measuring the weight of a product classified as a product requiring weight measurement based on a measured volume value among products passing through the transport means;
A sorting unit for classifying the product into any one suit among a plurality of suits according to the data received from the volume measurement unit or the weight measurement unit; including,
Commodity classification system.
13. The method of claim 12,

The weight measuring unit, characterized in that for measuring the weight of the product included in the volume range of the preset small and medium-sized product,
Commodity classification system.
14. The method of claim 13,
The sorting unit,
If the weight of the product measured by the weight measurement unit is greater than or equal to a preset value, characterized in that the control is performed so that the product is classified into a predetermined suit,
Commodity classification system.
15. The method of claim 14,
a sensing unit which is provided before the weight measurement unit and detects product entry into the weight measurement unit;
further comprising,
Commodity classification system.
13. The method of claim 12,
The weight measurement unit,
(t21) receiving a plurality of time series weight values measured by the weight measurement unit;
(t22) defining n weight values (n is an integer greater than or equal to 2) as a first window;
(t23) defining a maximum value within the first window as a first maximum value;
(t24) defining n weight values as the i-th window starting with the next-order weight value of the first weight value in the i-1th window (where i is an integer that increases according to the number of regressions starting from 2);
(t25) defining a maximum value within the ith window as an ith maximum value;
(t26) When the amount of change between the i-1 maximum value and the i-th maximum value is 0 or less, the i-1 maximum value is defined as a peak value, and when the amount of change is a value greater than 0 (t24) Characterized in defining the peak value by performing the step of regressing to the step,
At least one of the n weight values in the step (t24) after regression is characterized in that it has a lower priority in time series than the weight value of the first time in the i-th window,
product classification system

Images