Cargo handling command device and program

Abstract

To reduce the effects of reflections and glare that occur when imaging labels attached to packages wrapped in transparent resin film, and to acquire the management number written on the label with high accuracy, thereby accurately identifying the destination of the package. [Solution] The package transport command device 1 comprises a storage device 15 that stores transport instruction data associating the package's management number with the package's delivery destination, multiple cameras 21, 22, 23 that image the package from multiple directions and generate multiple images, a composite image generation unit (S21) that generates a composite image from the multiple images, an OCR processing unit (S22) that performs OCR processing on the composite image, a delivery destination identification unit (S24, S17) that queries the transport instruction data with the result of the OCR processing to identify the package's destination, and a transmission unit (S18) that transmits a transport command to the transport device to transport the package to the identified destination.

Description

Technical Field

[0001] This embodiment relates to a luggage transfer command device and a program that transmit a transfer command for transferring luggage to a transfer device.

Background Art

[0002] For example, various articles are carried into a factory. The articles are products, parts, raw materials, or semi-finished products, etc. that are targets for inspection, manufacturing, processing, assembly, or trading. Usually, the articles are carried in while being housed in cardboard boxes or the like.

[0003] Labels with management numbers (luggage numbers) for uniquely identifying the articles or luggage are pasted on the surfaces of many cardboard boxes. By using the management number, it becomes possible to efficiently execute various management processes such as inspection, storage, transfer, process management, and inventory management related to the articles.

[0004] In recent years, there are cases where the carried-in luggage is imaged with a camera, the character string of the management number is obtained by OCR, and it is utilized for transfer, process management, inventory management, etc. For this reason, if the management number cannot be accurately obtained, it may interfere with various processes in the factory, causing work stoppages and mis-transfers.

[0005] Here, the luggage is often carried in while being wrapped with a transparent resin film for the purpose of preventing collapse during transportation and preventing contamination. Such film packaging causes light reflection and refraction on the film surface, and when imaging the management number written on the label attached to the luggage with a camera, it causes halation and contrast reduction, resulting in OCR errors such as characters being misrecognized, characters not being recognized, or only a part of the character string being recognized when converting the character string included in the captured image into character data by OCR processing.

[0006] When such OCR errors occur, as mentioned above, they can cause delays in operations or misdeliveries, and can also lead to serious inventory management errors where the inventory information managed based on the management number does not match the actual quantity, storage location, or condition of the goods. [Overview of the project] [Problems that the invention aims to solve]

[0007] The objective is to reduce the effects of reflections and glare that occur when imaging labels attached to packages wrapped in transparent resin film, and to acquire strings of characters such as management numbers written on the labels with high accuracy, thereby accurately identifying the destination of the package. [Means for solving the problem]

[0008] An embodiment of the present invention provides a cargo transport command device for transporting a cargo having multiple labels affixed to multiple surfaces, each label bearing a string of characters for identifying the cargo, such as a management number. The device includes: a storage unit for storing transport instruction data that associates the destination of the cargo with the string of characters; multiple cameras for imaging the cargo from multiple directions and generating multiple images; a composite image generation unit for generating a composite image from the generated multiple images; an OCR processing unit for performing OCR processing on the composite image; a destination identification unit for identifying the destination of the cargo by querying the transport instruction data with the result of the OCR processing; and a transmission unit for transmitting a transport command to the transport device for transporting the cargo to the identified destination. [Brief explanation of the drawing]

[0009] [Figure 1] Figure 1 shows the configuration of a cargo transport command device according to one embodiment of the present invention. [Figure 2] Figure 2 shows an example of a transport instruction data table stored in the storage device (storage unit) shown in Figure 1. [Figure 3]Figure 3 is a perspective view showing the installation of the camera and photoelectric sensor shown in Figure 1, along with the cargo on the roller conveyor. [Figure 4] Figure 4 is a flowchart showing the procedure for identifying and issuing commands for transport destinations by executing the transport command program stored in the storage device (storage unit) shown in Figure 1. [Figure 5] Figure 5 schematically shows the flow of steps S13-S17 when step S16 in Figure 4 is "YES". [Figure 6] Figure 6 schematically shows the flow of processes S13, S20-S23, and S17 when process S16 in Figure 4 is "NO". [Figure 7] Figure 7 is a flowchart showing the procedure for identifying and issuing commands to the transport destination by executing a program, according to a second embodiment of the present invention. [Figure 8] Figure 8 schematically shows the flow of steps S13, S14, S31, S32, and S17 when step S16 in Figure 7 is "NO". [Modes for carrying out the invention]

[0010] Embodiments of the present invention will be described below with reference to the drawings. (First Embodiment) Figure 1 shows the configuration of a cargo transport command device 1 according to the first embodiment of the present invention. The cargo transport command device 1 transmits a transport command to a transport device 6 to transport cargo, for example, cargo brought into a factory, to a destination such as a storage location. Examples of transport devices 6 include automated guided vehicles (AGVs), forklift-type unmanned transport vehicles, and manned transport vehicles such as manned forklifts. Here, an automated guided vehicle (hereinafter referred to as AGV) will be used as an example to explain the transport device 6.

[0011] The cargo transport command device 1 consists of an information processing device (computer) 10, multiple cameras 21, 22, 23, a photoelectric sensor 24, and a wireless communication antenna 25. Here, we will explain using three cameras 21, 22, and 23 as an example, but any number of cameras, such as two or four or more, may be used.

[0012] The information processing device 10 has a processor 11. The processor 11 is equipped with RAM 13, ROM 14, storage device 15, interface (I / F) 16, short-range wireless communication unit 17, input devices 18 such as a keyboard or mouse, and a display 19 such as an LCD (Liquid Crystal Display), all connected via a data / control bus 12. The processor 11 is implemented by a CPU or the like. The RAM 13 functions as the main memory, work area, etc., of the processor 11. The ROM 14 stores the BIOS (Basic Input Output System), operating system program (OS), etc., which are executed by the processor 11.

[0013] The storage device 15 is implemented using a non-volatile storage medium such as an HDD or SSD. The storage device 15 stores transport instruction data, which associates a destination with a string of characters used to identify a package, such as a package management number, and also stores a program for executing the destination instruction processing according to this embodiment (hereinafter referred to as the transport instruction program). The transport instruction data is created in advance by the administrator before the package is brought in. Specifically, as illustrated in Figure 2, the transport instruction data is configured as a table that associates a package management number with a code that identifies the destination (destination code), destination type, priority, and remarks.

[0014] The string of characters used to identify a package, such as a management number, product number, name of the items contained in the package, delivery address code, lot number, etc., refers to any string of characters used to identify an item or package. For the sake of explanation, we will refer to it as a management number here.

[0015] The short-range wireless communication unit 17 is configured using a communication module compliant with the Wi-Fi (registered trademark) standard as an example of a wireless communication standard. The short-range wireless communication unit 17 transmits and receives data such as transport commands and location information to and from the AGV 6 via the wireless communication antenna 25, in accordance with the control of the processor 11.

[0016] As illustrated in Fig. 3(a), the package PA is carried in with the articles stored in a cardboard box or the like wrapped in a transparent resin film. Labels are attached to the upper surface and the four side surfaces of the cardboard box, respectively. Each label has a management number related to the articles stored in the cardboard box described thereon by printing or the like.

[0017] As shown in Figs. 3(b) and 3(c), the loading-side pallet stand 3 and the delivery-side pallet stand 4 are connected by a roller conveyor 2. The package PA placed on the pallet 7 is arranged on the loading-side pallet stand 3 by a manned forklift or the like and conveyed to the delivery-side pallet stand 4 via the roller conveyor 2. At the delivery-side pallet stand 4, the package PA is delivered to the AGV 6 together with the pallet 7. The AGV 6 is conveyed to each delivery destination according to the conveyance instruction received from the conveyance instruction device 1.

[0018] In front of the delivery-side pallet stand 4, a plurality of cameras 21, 22, 23 supported by a stand 5 are provided as an imaging unit for imaging the package PA moving on the roller conveyor 2 from a plurality of directions. The camera 21 images the package PA from above, the camera 22 images the package PA from the left side, and the camera 23 images the package PA from the right side. However, the imaging directions of the cameras 21, 22, 23 with respect to the package PA are not limited to above, left, and right.

[0019] A photoelectric sensor 24 is supported on the stand 5 together with the cameras 21, 22, 23. The photoelectric sensor 24 is selected from any of transmissive type, reflective type, diffuse reflection type, laser type, etc. The photoelectric sensor 24 detects the arrival of the package PA, and upon this detection, the plurality of cameras 21, 22, 23 image the package PA simultaneously.

[0020] The plurality of images simultaneously acquired by the plurality of cameras 21, 22, 23 are input to the processor 11. The processor 11 extracts a label area that is presumed to have a label attached to the package PA from the plurality of images, and generates a plurality of partial images. Since the label is white and the cardboard is light brown, the label area can be easily extracted based on the color difference between the two.

[0021] The processor 11 decomposes each partial image into a plurality of local areas, sets a weighting coefficient based on the luminance, contrast, edge strength, presence or absence of a reflection component, or other OCR reliability index of each local area, and weights and synthesizes the same local areas on the partial images according to the weighting coefficient, thereby generating a composite image in which the management number described on the label appears with high visibility.

[0022] The processor 11 executes OCR processing on the generated composite image. The OCR processing result for the composite image has a higher character recognition rate than the OCR processing result for the partial image. The character string as the OCR processing result represents the management number with high accuracy. The processor 11 queries the string as the OCR processing result against the conveyance instruction data stored in the storage device 15, and identifies the destination code of the package PA. The processor 11 transmits a conveyance command including the identified destination code to the AGV 6.

[0023] Referring to FIG. 4, the conveyance instruction procedure according to the present embodiment will be described in detail. FIG. 5 is a diagram schematically showing the flow of steps S13 - S17 when step S16 in FIG. 4 is "YES", and FIG. 6 is a diagram schematically showing the flow of steps S13, S20 - S23, S17 when step S16 in FIG. 4 is "NO".

[0024] The transport command procedure is implemented by the processor 11 executing a transport command program. The cargo PA brought into the factory is sequentially unloaded onto the pallet platform 3 on the receiving side by, for example, a manned forklift, and then moved along the roller conveyor 2 to the pallet platform 4 on the receiving side. The photoelectric sensor 24 waits to detect the cargo PA (S11). When the cargo PA reaches a position a predetermined distance before the pallet platform 4 on the receiving side, and the photoelectric sensor 24 detects the cargo PA (S11; YES), the processor 11 sends an imaging command to the multiple cameras 21, 22, and 23 (S12). In response to the imaging command, the cameras 21, 22, and 23 capture images of the cargo PA from above, the left side, and the right side, respectively. The imaging operations of the cameras 21, 22, and 23 generate images of the top surface, left side, and right side of the cargo PA.

[0025] The processor 11 extracts the label region from each of these three images and generates three sub-images relating to the label region (S13). As mentioned above, the background color of the label (also called the base material color or display color) is white or light-colored, and the color of the cardboard packaging is light brown, so the label region is extracted based on the color difference and brightness difference between the two.

[0026] Next, the processor 11 performs OCR (Optical Character Recognition) processing on each of the three partial images (S14). As is well known, OCR analyzes the shape of characters from image data and obtains the character (character code) corresponding to that shape. The recognition accuracy of the OCR process is affected by the contrast between the character and the background, the sharpness of the character, distortion, the presence or absence of reflection, etc.

[0027] OCR errors may occur, such as misrecognition of characters, missing characters, duplicate characters, or inability to recognize characters. When the exterior of the package PA is covered with a transparent film, the reflection or glare from the film reduces the contrast of the characters in the label image, making OCR errors more likely.

[0028] The processor 11 queries the transport instruction data for each of the three OCR processing results (strings) corresponding to the three partial images (S15), and extracts a management number consisting of the same string as each OCR processing result (string). Here, a management number consisting of the same string as the OCR processing result (string) means that each OCR processing result (string) and the management number have a complete match in terms of character type, number of characters, and character arrangement.

[0029] If a management number consisting of a string identical to the OCR processing result (string) is extracted from the transport instruction data, the OCR processing result is judged to be a correct string, and the management number can be judged to be correct as the management number for the cargo PA.

[0030] Then, if a management number consisting of the same string is extracted from at least one of the OCR processing results (strings) corresponding to the three partial images, the identification of the management number is considered successful, and step S16 is judged as YES. On the other hand, if a management number consisting of the same string is not extracted from all of the OCR processing results (strings) corresponding to the three partial images, the management number is considered not to exist, and step S16 is judged as NO.

[0031] Since it is sufficient for the OCR processing result obtained from at least one of the three images to match the management number, the success rate of identifying the management number can be improved compared to using a single image.

[0032] If the result in step S16 is determined to be YES, the processor 11 refers to the transport instruction data and obtains the destination code of the package PA associated with the management number (S17). The processor 11 generates a transport command and selects the AGV6 closest to the handover position from among the multiple AGV6s, and transmits the transport command to the AGV6 via the short-range wireless communication unit 17 (S18). The transport command includes at least the management number that identifies the package PA and the destination code that indicates the destination of the package PA. It also includes, as appropriate, the handover position, transport order or priority, transport route information, and the upper limit of the travel speed.

[0033] When the planned number of packages have been delivered (S19; YES), the process ends. If the number of delivered packages has not reached the planned number (S19; NO), the process returns to step S11 and waits for the arrival of the next package PA.

[0034] Processes S20 to S26 are executed when a NO determination is made in process S16, that is, when individual partial images are subjected to OCR processing and the transport instruction data is queryed from the OCR processing results, but it is not possible to extract a management number for all of the OCR processing results.

[0035] In step S20, the processor 11 performs at least one preprocessing step on the three partial images of the label area, such as brightness normalization, contrast normalization, scale normalization, and edge enhancement, with the aim of addressing film reflection, lighting unevenness, and blurring of characters.

[0036] Next, the processor 11 divides each of the three pre-processed partial images into multiple local regions, and generates a composite image by weighting and summing the pixel values ​​according to the weight coefficients determined for each local region (S21).

[0037] The weighting coefficients used to adjust the contribution of pixel values ​​in local regions are determined for each local region according to the OCR confidence index. The OCR confidence index is an index that indicates the probability that characters in an image will be correctly recognized, and it numerically represents the quality of the character image or the certainty of the character recognition result. For example, it is determined for each local region based on at least one of the following: contrast between characters and background, edge strength (clarity of character outlines), local brightness variation, and the ratio of reflections and blown-out highlights. Between corresponding local regions within the range of three partial images, local regions with higher OCR confidence indices are assigned larger coefficients for weighting, and local regions with lower OCR confidence indices are assigned smaller coefficients.

[0038] The weighting coefficients are determined by quantifying the contrast between the text and the background, the edge strength of the text outline, the variation in local brightness, and the percentage of reflection or blown-out highlights for each local region of each partial image, normalizing these values, and then integrating them to determine the weighting coefficients.

[0039] The contrast between characters and background is the difference in brightness between character pixels and background pixels within a local area; a larger value indicates higher character recognition accuracy. Edge strength (clarity of character outlines) is the average or integral value of the magnitude of the horizontal and vertical brightness gradients (edge ​​strength) calculated by the differential filter within a local area; a higher edge strength indicates clearer character outlines. Local brightness variability is the brightness dispersion or standard deviation within a local area; values ​​outside the appropriate range indicate a lot of noise or overexposure. The reflection / overexposure ratio is the proportion of high-brightness pixels above a predetermined threshold; a smaller ratio indicates less reflection, while a larger ratio indicates more overexposure.

[0040] These OCR confidence indices are normalized to a range of 0 to 1, and these normalized indices are weighted and added together using weighting coefficients predetermined according to the contribution of each indice to the OCR confidence index, thereby calculating weight coefficients for each local region.

[0041] In this way, partial images (label images) of the top, left, and right images are combined according to the OCR confidence index to reduce OCR errors and improve the OCR character recognition rate, and OCR processing is performed on the combined image.

[0042] In step S22, the processor 11 performs OCR processing on the composite image. Since the OCR confidence index of the composite image is increased for each local region, the OCR character recognition performance can be improved compared to subjecting each partial image to OCR processing individually.

[0043] The processor 11, similar to process S15, queries the transport instruction data for the OCR processing result (string) of the composite image (S23), and extracts a management number consisting of the same string as the OCR processing result (string).

[0044] If the identification of a management number consisting of a string that is exactly the same as the OCR processing result (string) of the composite image is successful (S24; YES), the process returns to step S17, and the processor 11 refers to the transport instruction data and obtains the transport destination code of the package PA associated with the management number. The processor 11 generates a transport command and transmits the transport command to the AGV 6 closest to the handover position via the short-range wireless communication unit 17 (S18). When the handover of the planned number of packages is completed (S19; YES), the process ends, and if the number of packages to be delivered has not reached the planned number (S19; NO), the process returns to step S11 and waits for the arrival of the next package PA.

[0045] In step S24, if the processor 11 fails to identify a management number consisting of a string that is exactly the same as the OCR processing result (string) for the composite image (S24; NO), it queries the transport instruction data with the OCR processing result (string) for the composite image, extracts multiple management numbers consisting of strings that partially match the OCR processing result (string), and displays a list of transport instruction data containing each of these multiple management numbers on the display 19 (S25).

[0046] The administrator visually inspects the multiple management numbers and other items of the transport instruction data displayed on the display 19, confirms the physical corresponding package PA, compares the package PA with the transport instruction data, and then operates the input device 18 to selectively specify one of the multiple management numbers (S26). Upon receiving the selection, the processor 11 identifies the destination based on the selected management number (S17) and transmits a transport command corresponding to that destination to the AGV6. Thus, even if the automated identification of the management number fails using composite images, the configuration allows for human intervention, thereby preventing delays in the transport operation.

[0047] Even when a single image and individual OCR processing fail to identify the management number, combining image synthesis, re-OCR processing of the synthesized image, and final verification by the administrator allows for highly reliable identification of the management number. (Second Embodiment) Figure 7 shows the procedure for identifying and commanding a transport destination by executing the transport command program according to the second embodiment of the present invention. Figure 8 schematically shows the flow of steps S13, S14, S31, S32, and S17 when step S16 in Figure 7 is "NO".

[0048] In the first embodiment, the system is configured to perform OCR processing after synthesizing images (partial images) acquired by multiple cameras 21, 22, and 23. In contrast, in this second embodiment, OCR processing is performed individually on the images acquired by each camera 21, 22, and 23, and a string merging process (step S31) is executed to merge the multiple strings obtained by the OCR processing. The merged string obtained by the merging process is then used to query transport instruction data (step S32).

[0049] This makes it possible to identify management numbers by utilizing the redundancy of character recognition results, even when image synthesis is difficult due to image reflection or distortion, thereby improving the accuracy of management number identification without relying on image synthesis processing.

[0050] The string merging process (step S31) involves normalizing the notation of multiple strings obtained as a result of OCR processing (step S14) on each image acquired by multiple cameras 21, 22, and 23, then mapping the character positions, and finally generating a merged string by selecting the correct character for each corresponding character position through a majority vote.

[0051] For notation normalization, since variations in notation occur between different cameras 21, 22, and 23, typical methods include unifying full-width / half-width characters, uppercase / lowercase letters, and the handling of spaces and symbols. Furthermore, for character position information mapping, since the arrangement of characters within the same label area matches, characters with the same position are aligned. After aligning character positions, the character that is typically assumed to be correct is selected for each character position by majority vote. When three characters are recognized at a given character position, the most frequently occurring character is selected as the canonical character. When two characters are recognized, the character with the higher character recognition confidence is selected as the canonical character. Character recognition confidence is an indicator of the degree to which the character recognized by OCR processing is estimated to be correct, and can be determined, for example, by the contrast between the character area and the background in the original partial image, the clarity of the edges of the character area, etc. When only one character is recognized at a given character position, that character is naturally determined as the canonical character.

[0052] According to this second embodiment, OCR processing is performed individually on each image acquired by the multiple cameras 21, 22, and 23, and the transport instruction data is queried based on a unified string obtained by integrating the multiple strings obtained as OCR results. Therefore, even if some OCR results contain errors due to image reflection or distortion, the management number can be identified with high accuracy. This improves the rate of automatic identification of management numbers, reduces the frequency of manual intervention, and suppresses delays in transport operations.

[0053] In the above example, multiple images of the package PA, captured simultaneously from multiple directions, were combined and subjected to OCR processing. However, the label of the package PA may also be captured from multiple images repeatedly from only one direction, such as above, to the right, or to the left, and then combined and subjected to OCR processing. Furthermore, while a management number was obtained by combining multiple images of the package PA, captured repeatedly from one direction, and subjected to OCR processing, the method is not limited to a management number and may use other strings of characters. Moreover, the target is not limited to package PA, but may also be other objects such as goods. Of course, it is not limited to transport instructions and may be applied to other situations such as goods transactions. This embodiment is not limited to a package transport instruction device that utilizes OCR character recognition results for package transport instructions. The OCR character recognition results may be used for purposes other than package transport, and the use of the OCR character recognition results may be applied simply to an OCR character recognition processing device without limiting the use to a specific purpose.

[0054] It should be noted that the embodiments are not limited to those described above, and various modifications can be made during implementation without departing from the gist of the invention. Furthermore, each embodiment may be combined as appropriate, and in that case, the combined effects can be obtained. Moreover, the embodiments described above include various inventions, and various inventions can be extracted by selecting combinations from the multiple constituent elements disclosed. For example, if the problem can be solved and effects obtained even if some constituent elements are deleted from all the constituent elements shown in the embodiment, then the configuration with these deleted constituent elements can be extracted as an invention. [Explanation of Symbols]

[0055] 1...Cargo handling command device, 2...Roller conveyor, 3,4...Pallet stand, 5...Stand, 6...Automated guided vehicle (AGV), 7...Pallet, 10...Information processing device, 11...Processor, 12...Data / control bus, 13...RAM, 14...ROM, 15...Storage device, 16...Interface, 17...Short-range wireless communication unit, 21...Upper camera, 22...Left camera, 23...Right camera, 24...Photoelectric sensor, 25...Wireless communication antenna, PA...Cargo.

Claims

1. A package transport command device that commands a transport device to transport a package, wherein multiple labels are affixed to multiple surfaces, each of which has a string of characters for identifying the package, such as a management number, A storage unit that stores transport instruction data associating the delivery destination of the package with the aforementioned string, Multiple cameras capture images of the aforementioned cargo from multiple directions and generate multiple images, A composite image generation unit generates a composite image from the multiple images generated, An OCR processing unit that performs OCR processing on the composite image, A destination identification unit identifies the destination of the cargo by querying the transport instruction data for the result of the OCR processing, A cargo transport command device comprising: a transmitting unit that transmits a transport command to the transport device for transporting the cargo to the specified destination.

2. The system further includes an extraction unit that extracts multiple partial images relating to the label region from the multiple images, The cargo transport command device according to claim 1, wherein the composite image generation unit generates the composite image by combining the plurality of partial images.

3. The luggage transport command device according to claim 1, further comprising a preprocessing unit that performs preprocessing on the plurality of partial images, including at least one of luminance normalization, contrast normalization, scale normalization, and edge enhancement.

4. The cargo transport command device according to claim 2, wherein the composite image generation unit divides each of the plurality of partial images into a plurality of local regions, and generates the composite image by weighting and adding the plurality of partial images for each local region based on brightness, contrast, edge intensity, or reliability in the OCR processing.

5. The system further includes a sensor for detecting the arrival of the aforementioned package. The luggage transport command device according to claim 1, wherein the plurality of cameras simultaneously capture images of the luggage based on the output of the sensors.

6. The cargo transport command device according to claim 1, further comprising a management number extraction unit that, when the destination identification unit cannot identify the destination of the cargo by querying the transport instruction data for the result of the OCR processing, extracts a plurality of candidate management numbers that partially match the result of the OCR processing and displays them in a list.

7. A package transport command device that commands a transport device to transport a package, wherein multiple labels are affixed to multiple surfaces, each of which has a string of characters for identifying the package, such as a management number, A storage unit that stores transport instruction data associating the delivery destination of the package with the aforementioned string, Multiple cameras capture images of the aforementioned cargo from multiple directions and generate multiple images, An OCR processing unit that performs OCR processing on each of the multiple images generated, An integration processing unit that integrates the results of the OCR processing, A destination identification unit identifies the destination of the cargo by querying the transport instruction data for the results of the integrated OCR processing, A cargo transport command device comprising: a transmitting unit that transmits a transport command to the transport device for transporting the cargo to the specified destination.

8. An information processing device that commands a transport device to transport a package, wherein multiple labels are affixed to multiple sides, and each of the labels has a string of characters for identifying the package, such as a management number. A means for generating a composite image from multiple images obtained by imaging the aforementioned cargo from multiple directions, Means for performing OCR processing on the composite image, A means for identifying the destination of the package by querying the result of the OCR processing against transport instruction data that associates the destination of the package with the string, A program that functions as a means for transmitting a transport command to the transport device for transporting the cargo to the specified destination.

9. An information processing device that commands a transport device to transport a package, wherein multiple labels are affixed to multiple sides, and each of the labels has a string of characters for identifying the package, such as a management number. A means for separately performing OCR processing on multiple images generated by imaging the aforementioned cargo from multiple directions, Means for integrating the results of the OCR processing, A means for identifying the destination of the package by querying the result of the integrated OCR processing against transport instruction data that associates the string with the destination of the package, A program that functions as a means for transmitting a transport command to the transport device for transporting the cargo to the specified destination.

Images