WO2017087244A1 - Procédés et systèmes d'estimation de remplissage de conteneur - Google Patents

Procédés et systèmes d'estimation de remplissage de conteneur Download PDF

Info

Publication number
WO2017087244A1
WO2017087244A1 PCT/US2016/061279 US2016061279W WO2017087244A1 WO 2017087244 A1 WO2017087244 A1 WO 2017087244A1 US 2016061279 W US2016061279 W US 2016061279W WO 2017087244 A1 WO2017087244 A1 WO 2017087244A1
Authority
WO
WIPO (PCT)
Prior art keywords
grid
depth
container
shipping container
point cloud
Prior art date
Application number
PCT/US2016/061279
Other languages
English (en)
Inventor
Yan Zhang
Jay J. Williams
Kevin J. O'CONNELL
Cuneyt M. TASKIRAN
Original Assignee
Symbol Technologies, Llc
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from US14/944,860 external-priority patent/US9940730B2/en
Priority claimed from US14/978,367 external-priority patent/US10713610B2/en
Application filed by Symbol Technologies, Llc filed Critical Symbol Technologies, Llc
Priority to PL430920A priority Critical patent/PL239620B1/pl
Priority to CA3005452A priority patent/CA3005452C/fr
Priority to PL426752A priority patent/PL426752A1/pl
Priority to GB1807994.7A priority patent/GB2558507B/en
Priority to MX2018006105A priority patent/MX2018006105A/es
Priority to DE112016005287.1T priority patent/DE112016005287T5/de
Publication of WO2017087244A1 publication Critical patent/WO2017087244A1/fr

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0004—Industrial image inspection
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01F—MEASURING VOLUME, VOLUME FLOW, MASS FLOW OR LIQUID LEVEL; METERING BY VOLUME
    • G01F17/00—Methods or apparatus for determining the capacity of containers or cavities, or the volume of solid bodies
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/77—Retouching; Inpainting; Scratch removal
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/50—Depth or shape recovery
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/60—Analysis of geometric attributes
    • G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2200/00—Indexing scheme for image data processing or generation, in general
    • G06T2200/04—Indexing scheme for image data processing or generation, in general involving 3D image data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10016—Video; Image sequence
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10028—Range image; Depth image; 3D point clouds
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20021—Dividing image into blocks, subimages or windows
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20036—Morphological image processing

Definitions

  • Efficient loading of containers is a key element to successful distribution in the transportation and logistics industry. Ensuring that each container is loaded efficiently throughout the loading process is vital to successful distribution. However, the inability to verify that each container meets this goal has been a problem in the industry.
  • FIG. 1 depicts a shipping container, in accordance with some embodiments.
  • FIG. 2A depicts a flat back surface of a shipping container, in accordance with some embodiments.
  • FIG. 2B depicts a curved back surface of a shipping container, in accordance with some embodiments.
  • FIG. 3 depicts a loaded-container point cloud, in accordance with some embodiments.
  • FIG. 4 depicts a segmented loaded-container point cloud, in accordance with some embodiments.
  • FIG. 5 depicts an expanded-grid-element view of a segmented loaded- container point cloud, in accordance with some embodiments.
  • FIG. 6 depicts an architectural view of an example computing device, in accordance with some embodiments.
  • FIG. 7 depicts a first example method, in accordance with some embodiments.
  • FIG. 8 depicts a shipping container having an optically readable identifier, in accordance with some embodiments.
  • FIG. 9 depicts a second example method, in accordance with some embodiments.
  • FIG. 10 depicts an example scenario for detecting occlusions in a shipping container, in accordance with some embodiments.
  • FIG. 11 depicts an example sub-process for detecting close occlusions, in accordance with some embodiments.
  • FIG. 12 depicts an example sub-process for detecting far occlusions, in accordance with some embodiments.
  • FIG. 13 depicts an example of temporal analysis, in accordance with some embodiments.
  • FIGs. 14A and 14B depict examples of graphed container-fullness- estimation results without and with occlusion correction, respectively, in accordance with some embodiments.
  • One embodiment takes the form of a process that includes (a) receiving a three-dimensional (3D) point cloud from a depth sensor that is oriented towards an open end of a shipping container, where the point cloud includes a plurality of points that each have a respective depth value, (b) segmenting the received 3D point cloud among a plurality of grid elements, (c) calculating a respective loaded-container-portion grid- element volume for each grid element, (d) calculating a loaded-container-portion volume of the shipping container by aggregating the calculated respective loaded- container-portion grid-element volumes, (e) calculating an estimated fullness of the shipping container based on the loaded-container-portion volume and a capacity of the shipping container; and (f) outputting the calculated estimated fullness of the shipping container.
  • 3D three-dimensional
  • Another embodiment takes the form of a system that includes a depth sensor, a communication interface, a processor, a data storage containing instructions executable by the processor for causing he system to carry out at least the functions described in the preceding paragraph.
  • the plurality of grid elements collectively forms a two-dimensional (2D) grid image that corresponds to a plane that is parallel to the open end of the shipping container, where each grid element has a respective grid- element area
  • the method further includes determining a respective loaded- container-portion grid-element depth value for each grid element, where calculating the respective loaded-container-portion grid-element volume for each grid element is based on at least the respective grid-element area and the respective loaded-container-portion grid-element depth value for each respective grid element.
  • the method further includes determining an unloaded-container-portion depth value for each grid element, and determining a respective loaded-container-portion grid-element depth value for each grid element is based at least in part on the difference between (i) a depth dimension of the shipping container and (ii) the determined unloaded-container-portion depth value for the corresponding grid element.
  • assigning the grid-element depth value for the given grid element based on the depth values of the points in the point cloud that correspond to the given grid element includes assigning as the grid-element depth value for the given grid element a minimum value from among the depth values of the points in the point cloud that correspond to the given grid element. [0027] In at least one embodiment, assigning the grid-element depth value for the given grid element based on the depth values of the points in the point cloud that correspond to the given grid element includes assigning as the grid-element depth value for the given grid element an average value of the depth values of the points in the point cloud that correspond to the given grid element.
  • the depth dimension of the shipping container is a grid-element-specific depth dimension that is based on a corresponding grid element in a reference empty-container point cloud.
  • the reference empty-container point cloud reflects a back wall of the shipping container being a flat surface. In at least one other such embodiment, the reference empty- container point cloud reflects a back wall of the shipping container being a curved surface.
  • the method further includes cleaning up the 2D grid image prior to determining a respective loaded-container-portion grid-element depth value for each grid element.
  • the depth sensor has an optical axis and an image plane
  • the method further includes, prior to segmenting the received point cloud among the plurality of grid elements, rotating the received 3D point cloud to align (i) the optical axis with a ground level and (ii) the image plane with an end plane of the shipping container.
  • rotating the point cloud is based on an offline calibration process using the ground level and the end plane as reference.
  • the method further includes determining the capacity of the shipping container based at least in part on the received 3D point cloud.
  • the method further includes (i) receiving an optical image of the shipping container and (ii) determining the capacity of the shipping container based at least in part on the received optical image.
  • determining the capacity of the shipping container based at least in part on the received optical image includes (i) determining at least one physical dimension of the shipping container from the received optical image and (ii) determining the capacity of the shipping container based on the at least one determined physical dimension.
  • determining the capacity of the shipping container based at least in part on the received optical image includes (i) using optical character recognition (OCR) on the at least one received optical image to ascertain at least one identifier of the shipping container and (ii) using the at least one ascertained identifier of the shipping container to determine the capacity of the shipping container.
  • OCR optical character recognition
  • each grid element has sides substantially equal to 5 millimeters (mm) in length.
  • each grid element is substantially square in shape, and a grid-element side length is an adjustable parameter.
  • One embodiment takes the form of a method that includes receiving a depth frame from a depth sensor oriented towards an open end of a shipping container, where the depth frame is projected to a 2D grid map which includes a plurality of grid elements that each have a respective depth value; identifying one or more occlusions in the depth frame; correcting the one or more occlusions in the depth frame using one or more temporally proximate depth frames; and outputting the corrected depth frame for fullness estimation.
  • the one or more occlusions includes a missing-data occlusion.
  • identifying the missing-data occlusion includes (i) generating a binarization map delineating between (a) grid elements for which the respective depth value is valid and (b) grid elements for which the respective depth value is not valid and (ii) identifying the missing-data occlusion as a cluster of grid elements in the binarization map for which the respective depth value is not valid.
  • identifying the missing-data occlusion further includes confirming that the identified cluster of grid elements exceeds a predetermined occlusion-size threshold.
  • identifying the missing-data occlusion further includes performing edge detection on the cluster of grid elements.
  • identifying the missing-data occlusion further includes performing contour identification on the cluster of grid elements.
  • the one or more occlusions includes a moving occlusion.
  • the moving occlusion is associated with a single grid element in the plurality of grid elements.
  • identifying the moving occlusion includes identifying a threshold depth change in the single grid element between the depth frame and at least one temporally proximate depth frame.
  • identifying the moving occlusion includes identifying that the depth value associated with the single grid element decreases with respect to previous frames and then increases in succeeding frames in less than a threshold amount of time across multiple depth frames.
  • the one or more occlusions includes a discontinuous occlusion.
  • identifying the discontinuous occlusion includes identifying a cluster of grid elements having a collective depth value that is more than a threshold difference less than a depth value of a loaded-portion boundary of the shipping container.
  • identifying the discontinuous occlusion includes confirming that the identified cluster of grid elements exceeds a predetermined occlusion-size threshold.
  • identifying the discontinuous occlusion further includes performing edge detection on the cluster of grid elements.
  • identifying the discontinuous occlusion further includes performing contour identification on the cluster of grid elements.
  • the grid elements are pixels. In some embodiments, the grid elements are groups of pixels.
  • the one or more identified occlusions corresponds to an occlusion set of the grid elements in the depth frame
  • correcting the one or more occlusions in the depth frame using one or more temporally proximate depth frames includes overwriting the occlusion set in the depth frame with data from corresponding non-occluded grid elements from one or more of the temporally proximate depth frames.
  • identifying the one or more occlusions includes analyzing a buffer of depth frames, where the buffer includes the received depth frame.
  • One embodiment takes the form of a system that includes a depth sensor oriented towards an open end of a shipping container, a communication interface, a processor, and data storage containing instructions executable by the processor for causing the system to carry out a set of functions, where the set of functions includes: receiving a depth frame from the depth sensor, where the depth frame includes a plurality of grid elements that each have a respective depth value; identifying one or more occlusions in the depth frame; correcting the one or more occlusions in the depth frame using one or more temporally proximate depth frames; and outputting the corrected depth frame for fullness estimation.
  • any of the variations and permutations described herein can be implemented with respect to any embodiments, including with respect to any method embodiments and with respect to any system embodiments. Furthermore, this flexibility and cross-applicability of embodiments is present in spite of the use of slightly different language (e.g., process, method, steps, functions, set of functions, and the like) to describe and or characterize such embodiments.
  • FIG. 1 depicts a shipping container, in accordance with some embodiments.
  • FIG. 1 depicts (i) a shipping container 102 and (ii) a depth sensor 104 that is oriented towards an open end of the shipping container 102.
  • the shipping container 102 could be designed for travel by truck, rail, boat, plane, and/or any other suitable mode or modes of travel.
  • the shipping container 102 could have any of a number of different shapes; a substantially rectangular shape (i.e., a rectangular cylinder) is depicted by way of example in FIG. 1.
  • the shipping container 102 contains objects (e.g., boxes and/or other packages) 106.
  • the shipping container 102 may have a number of different surfaces, perhaps flat, perhaps curved, among numerous other possibilities that could be listed here.
  • depth sensor 104 there are a number of types of depth sensor 104 that could be used, perhaps one that includes an RGB sensor, perhaps leap motion, perhaps Intel perceptual computing, perhaps Microsoft Kinect, among numerous other possibilities that could be listed here. There are also a number of depth-sensing techniques that could be implemented by the depth sensor 104, perhaps using stereo triangulation, perhaps using time of flight, perhaps using coded aperture, among numerous other possibilities that could be listed here.
  • the depth sensor 104 could be mounted to a wall or column or the like in a given shipping warehouse, and the shipping container 102 could be positioned on the back of a truck, and then driven (e.g., backed) into a position such that the depth sensor 104 is oriented towards an open end of the shipping container 102, as is depicted in FIG. 1.
  • FIGs. 2A and 2B Two examples are shown in FIGs. 2A and 2B.
  • FIG. 2A depicts (i) a flat back wall (i.e., surface) 202 of a shipping container and (ii) a depth sensor 204
  • FIG. 2B depicts (i) a curved back wall (i.e., surface) 206 of a shipping container and (ii) a depth sensor 208.
  • FIG. 2A depicts (i) a flat back wall (i.e., surface) 202 of a shipping container and (ii) a depth sensor 204
  • FIG. 2B depicts (i) a curved back wall (i.e., surface) 206 of a shipping container and (ii) a depth sensor 208.
  • numerous other examples of shipping-container shapes could be presented here.
  • FIG. 3 depicts a loaded-container point cloud, in accordance with some embodiments.
  • FIG 3 depicts a 3D point cloud 302.
  • the depth sensor that is oriented at the open end of the shipping container may gather depth information in a given field of view and transmit that information to a system that may be equipped, programmed, and configured to carry out the present systems and methods.
  • That set of information i.e., points
  • That set of information is referred to herein as being a 3D point cloud (or at times simply a point cloud); each point in such a cloud corresponds to a perceived depth at a corresponding point in the field of view of the depth sensor.
  • an outline 304 of a shipping container is shown, as are outlines 306A, 306B, and 306C of example packages in the example shipping container. These outlines 304 and 306A-C are intended to generally correspond to the shipping container 104 and the packages 106 that are depicted in FIG. 1, in order to help the reader to visualize an example real-world scenario from which the example point cloud 302 could have been derived, gathered, or the like.
  • each point in the point cloud 302 is shown in FIG. 3 as having an integer number that corresponds to an example depth value (in, e.g., example units such as meters). In actual implementations, any number of points could be present in the point cloud 302, as the various points that are depicted in FIG. 3 as being part of the point cloud 302 are for illustration and are not meant to be comprehensive.
  • the depth sensor that is oriented towards an open end of the shipping container has a vantage point with respect to the open end of the shipping container that is not aligned with the center of the open end of the shipping container in one or more dimensions. That is, the depth sensor and the shipping container might be relatively positioned such that the depth sensor is looking to some extent from one side or the other and could be vertically off center (e.g., elevated) as well. So, for example, the depth sensor may be positioned higher and to the right of the center of the plane that corresponds with the open end of the shipping container.
  • the present disclosure includes segmentation and projection of the received point cloud into a number of grid elements in a 2D grid map that collectively correspond to the open end of the shipping container.
  • this segmentation and projection step can be proceeded to without first performing one or more geometric rotations.
  • the present systems and methods include a step of one or more geometric rotations in accordance with the relative positions of the depth sensor and the open end of the shipping container. Such relative position can be pre-programmed into the system, or could otherwise be determined using depth sensors, optical cameras, and/or other suitable equipment.
  • FIG. 4 depicts a segmented loaded-container point cloud, in accordance with some embodiments.
  • FIG. 4 depicts a segmented 3D point cloud 402, which may be generated in a number of different ways, such as edge-based segmentation, surfaced-based segmentation, and/or scanline-based segmentation, among numerous other possibilities that may be listed here.
  • FIG. 4 depicts the segmented point cloud 402 after any necessary rotations were performed to account for the relative positions and alignments of the depth sensor and the open end of the shipping container.
  • the point cloud 402 is segmented among a plurality of grid elements, which collectively form a 2D grid image that corresponds to a plane that is parallel to the open end of the shipping container.
  • Each grid element has a respective grid-element area.
  • the grid elements are shown as being substantially square (e.g., 5 mm by 5 mm), though this is by way of example and not limitation, as any suitable dimensions and/or shapes could be used as deemed suitable by those of skill in the art for a given implementation.
  • the side length of the grid elements is an adjustable parameter. In some cases, this parameter is set to be as small a value as the associated depth sensor allows and/or is capable of.
  • one example grid element 404 is highlighted by way of example.
  • the grid element 404 is depicted as including ten total points from the segmented point cloud 402; four of those ten points have a depth value of 1 (e.g., 1 meter), five of those ten points have a depth value of 2 (e.g., 2 meters), and one of those ten points has a depth value of 3 (e.g., 3 meters).
  • This number of points in grid element 404 and these respective depth values are provided purely by way of example and for illustration, and in no way for limitation.
  • FIG. 5 depicts an expanded-grid-element view of a segmented loaded- container point cloud, in accordance with some embodiments.
  • FIG. 5 depicts a segmented 3D point cloud 502 (though zoomed out too far to depict individual points) and an expanded grid element 504.
  • the expanded grid element 504 includes, by way of example only, the same set of ten points that are in the grid element 404 of FIG. 4, albeit in a different arrangement; i.e., there are ten total points, including four points having a depth value of 1, five points having a depth value of 2, and one point having a depth value of 3.
  • the grid element 504 is assigned a characteristic depth value based on the depth values of the points in the subsection of the 3D point cloud that is found in the particular grid element 504. From among those depth values, the characteristic depth value for the grid element could be a minimum value, a mode (i.e., most commonly occurring) value, an average value, or some other possibility. Using the example data that is present in FIG. 5: if the minimum value were used, then the characteristic depth value for the grid element 504 would be 1; if the mode value were used, then the characteristic depth value for the grid element 504 would be 2; if the average value were used, then the characteristic depth value for the grid element 504 would be 1.7 (or 2 if rounded to the nearest whole number). And certainly numerous other possible implementations could be listed here. As is described more fully below, the characteristic depth value that is assigned to a given grid element is then used, along with the area of that grid element, to calculate a loaded-portion volume for that particular grid element.
  • FIG. 6 depicts an architectural view of an example computing device, in accordance with some embodiments.
  • the example computing device 600 may be configured to carry out the functions described herein, and as depicted includes a communications interface 602, a processor 604, data storage 606 (that contains program instructions 608 and operational data 610), a user interface 612, peripherals 614, and a communication bus 616.
  • a communications interface 602 includes a processor 604, data storage 606 (that contains program instructions 608 and operational data 610), a user interface 612, peripherals 614, and a communication bus 616.
  • the communication interface 602 may be configured to be operable for communication according to one or more wireless-communication protocols, some examples of which include LMR, LTE, APCO P25, ETSI DMR, TETRA, Wi-Fi, Bluetooth, and the like.
  • the communication interface 602 may also or instead include one or more wired-communication interfaces (for communication according to, e.g., Ethernet, USB, and/or one or more other protocols.)
  • the communication interface 602 may include any necessary hardware (e.g., chipsets, antennas, Ethernet interfaces, etc.), any necessary firmware, and any necessary software for conducting one or more forms of communication with one or more other entities as described herein.
  • the processor 604 may include one or more processors of any type deemed suitable by those of skill in the relevant art, some examples including a general -purpose microprocessor and a dedicated digital signal processor (DSP).
  • processors of any type deemed suitable by those of skill in the relevant art, some examples including a general -purpose microprocessor and a dedicated digital signal processor (DSP).
  • DSP dedicated digital signal processor
  • the data storage 606 may take the form of any non-transitory computer- readable medium or combination of such media, some examples including flash memory, read-only memory (ROM), and random-access memory (RAM) to name but a few, as any one or more types of non-transitory data-storage technology deemed suitable by those of skill in the relevant art could be used.
  • the data storage 606 contains program instructions 608 executable by the processor 604 for carrying out various functions described herein, and further is depicted as containing operational data 610, which may include any one or more data values stored by and/or accessed by the computing device in carrying out one or more of the functions described herein.
  • the user interface 612 may include one or more input devices (a.k.a. components and the like) and/or one or more output devices (a.k.a. components and the like.) With respect to input devices, the user interface 612 may include one or more touchscreens, buttons, switches, microphones, and the like. With respect to output devices, the user interface 612 may include one or more displays, speakers, light emitting diodes (LEDs), and the like. Moreover, one or more components (e.g., an interactive touchscreen and display) of the user interface 612 could provide both user- input and user-output functionality.
  • the peripherals 614 may include any computing device accessory, component, or the like, that is accessible to and useable by the computing device 600 during operation.
  • the peripherals 614 includes a depth sensor.
  • the peripherals 614 includes a camera for capturing digital video and/or still images. And certainly other example peripherals could be listed.
  • FIG. 7 depicts a first example method, in accordance with some embodiments.
  • FIG. 7 depicts a method 700 that includes steps 702, 704, 706, 708, 710, and 712, and is described below by way of example as being carried out by the computing system 600 of FIG. 6, though in general the method 700 could be carried out by any computing device that is suitably equipped, programmed, and configured.
  • the computing system 600 receives a 3D point cloud from a depth sensor that is oriented towards an open end of a shipping container.
  • the point cloud includes a plurality of points that each have a respective depth value.
  • the computing system 600 upon receiving the 3D point cloud, may rotate the received 3D point cloud to align (i) an optical axis of the depth sensor with a ground level and (ii) an image plane of the depth sensor with an end plane of the shipping container.
  • This rotating of the received point cloud may be based on a calibration process (e.g., an offline calibration process) that uses the ground level and the end plane as reference.
  • the computing system 600 segments the 3D point cloud that was received at step 702 among a plurality of grid elements.
  • those grid elements could be substantially rectangular (e.g., square) in shape, and they may collectively form a 2D grid image that corresponds to a plane that is parallel to the open end of the shipping container, where each grid element has a respective grid-element area.
  • the computing system 600 calculates a respective loaded- container-portion grid-element volume for each grid element.
  • the computing system 600 may do so by first determining a respective loaded-container-portion grid-element depth value for each grid element, and then determining each respective loaded- container-portion grid-element volume for each grid element by multiplying the particular grid element's area by the particular grid element's respective loaded- container-portion grid-element depth value.
  • the computing system 600 cleans up the 2D grid image prior to determining a respective loaded- container-portion grid-element depth value for each grid element.
  • the computing system 600 may determine a particular grid element's respective loaded-container-portion grid-element depth value, in one embodiment the computing system 600 determines an unloaded-container-portion depth value for the particular grid element, and then determines the respective loaded- container-portion grid-element depth value for the particular grid element based at least in part on the difference between (i) a depth dimension of the shipping container and (ii) the determined unloaded-container-portion depth value for the corresponding grid element.
  • the computing system 600 could determine that the loaded-container-portion depth value for the given grid element was 47 meters.
  • the computing system 600 may determine the unloaded- container-portion depth value for a given grid element.
  • the computing system 600 assigns a characteristic grid-element depth value to the given grid element based on the depth values of the points in the point cloud that correspond to the given grid element. As described above, some options for doing so including selecting a minimum value, a mode value, and an average value. A maximum value could also be selected, though this would tend to lead to underloading of containers by overestimating their fullness, which would be less than optimally efficient.
  • the computing system 600 may then determine the respective unloaded- container-portion depth value for the given grid element based at least in part on the difference between (i) the assigned characteristic grid-element depth value for the given grid element and (ii) an offset depth value corresponding to a depth between the 3D depth sensor and a front plane of the shipping container.
  • the computing system 600 may consider the unloaded-container-portion depth value for that grid element to be 3 meters. And certainly numerous other examples could be listed.
  • the depth dimension of the shipping container that is used to derive a loaded-container-portion depth value from an unloaded-container-portion depth value for a given grid element is a grid-element-specific depth dimension that is based on a corresponding grid element in a reference empty-container point cloud.
  • the back wall could be flat or curved, as depicted in FIGs. 2A and 2B, and the grid-element-specific depth dimension for a given grid element could accordingly reflect this.
  • a reference point cloud could be gathered using an empty shipping container of the same type, and that reference point cloud could be stored in data storage and recalled, perhaps on a grid-element-by-grid-element basis to perform the herein-described calculations.
  • the computing system 600 calculates a loaded-container- portion volume of the shipping container by aggregating the respective loaded- container-portion grid-element volumes that were calculated at step 706, giving a result that corresponds to what volume (in, e.g., cubic meters) of the shipping container has been loaded. It is noted that loaded in this context essentially means no longer available for loading. Thus, empty space that is now inaccessible due to packages being stacked in the way would be counted as loaded right along with space in the shipping container that is actually occupied by a given package.
  • the computing system 600 calculates an estimated fullness of the shipping container based on (i) the loaded-container-portion volume that was calculated at step 708 and (ii) a capacity of the shipping container.
  • the estimated fullness of the shipping container may be calculated as the loaded-portion volume of the shipping container divided by the capacity of the shipping container.
  • the capacity of the shipping container could be determined in multiple different ways, some of which are described below.
  • the computing system 600 determines the capacity of the shipping container based at least in part on the received 3D point cloud.
  • the 3D point cloud may be indicative of the dimensions of the shipping container such that the capacity of the shipping container can be determined.
  • the computing system receives an optical image of the shipping container, and determines the capacity of the shipping container based at least in part on the received optical image. This could include determining actual dimensions of the shipping container from the optical image, and could instead or in addition include extracting an identifier of the shipping container from the optical image, perhaps using optical character recognition (OCR), and then querying a local or remote database using that identifier in order to retrieve dimension and/or capacity data pertaining to the particular shipping container.
  • OCR optical character recognition
  • the system may determine that the entire interior of the shipping container is not visible to the depth sensor, perhaps due to the relative location and arrangement of the depth sensor and the shipping container.
  • the system may define a volume of interest (VOI) as being the part of the interior of the container that is visible to the depth sensor.
  • the system may in some such instances calculate the estimated fullness of the container to be loaded portion of the VOI divided by the capacity (i.e., total volume) of the VOI.
  • the system may simply assume that any internal portion of the shipping container that cannot be seen with the depth camera is loaded, and in such cases may still calculate the estimated fullness as the loaded portion of the entire shipping container divided by the total capacity of the entire shipping container.
  • the computing system 600 outputs the calculated estimated fullness of the shipping container, perhaps to a display, perhaps to a data storage, perhaps using wireless and/or wired communication to transmit the calculated estimated fullness of the shipping container to one or more other devices or systems, and/or perhaps to one or more other destinations.
  • FIG. 8 depicts a shipping container having an optically readable identifier in accordance with some embodiments.
  • FIG. 8 depicts a container 802, an indicia 804 (e.g., bar code or alphanumeric identifier), and an optical reader 806.
  • an optical reader 806 acquires an alphanumeric identifier of the container data using OCR.
  • the computing system may then use that acquired alphanumeric identifier of the container to query a database for dimension data pertaining to the shipping container.
  • other example implementations could be listed here as well.
  • occlusions there may be one or more moving or stationary occlusions (e.g., package loaders, stray packages, etc.) between the 3D depth sensor and the loaded portion of the container.
  • Some occlusions cause underestimates of container fullness, perhaps by being so close to the 3D depth sensor so as to create gaps in the point-cloud data.
  • Some occlusions cause overestimates of container fullness, perhaps by being so close to actually loaded packages so as to be confused for (e.g., clustered with) those loaded packages.
  • the presence of occlusions can result in erroneous estimation of the fullness of the shipping container.
  • FIG. 9 depicts a second example method, in accordance with some embodiments.
  • a method 900 which includes the steps of receiving, at step 902, a depth frame from a depth sensor oriented towards an open end of a shipping container, where the depth frame includes a plurality of grid elements that each have a respective depth value.
  • the method 900 further includes identifying, at step 904, one or more occlusions in the depth frame. In some instances, only one or more far occlusions are detected. In some instances, only one or more close occlusions are detected. In some instances, both far and close occlusions are detected.
  • the method 900 further includes correcting, a step 906, the one or more occlusions in the depth frame using one or more temporally proximate depth frames, and outputting, at step 908, the corrected depth frame for fullness estimation.
  • FIG. 10 depicts an example scenario for detecting occlusions in a shipping container, in accordance with some embodiments.
  • FIG. 10 depicts an example scenario in which a depth sensor 1030 is configured to collect depth data while oriented towards a shipping container 1000.
  • close occlusions maybe caused by occluding objects close to depth sensor 1030 (e.g., loaders, unloaded packages, see FIG. 10 object 1005).
  • close occlusions appear as gaps or holes (no data (or no valid data)) in the 3D depth data, which may result in underestimated fullness, as less than a complete set of 3D depth volume data is processed.
  • close occlusions often present as gaps in data, they may also be referred to as "missing-data occlusions.”
  • an object 1005 is within the depth sensor's minimum detection range, and the depth sensor may therefore not provide any data for areas blocked by object 1005.
  • This gap in the 3D depth data may cause the system to omit the volume occupied by region 1010 while calculating fullness, thus resulting in an underestimate of the shipping-container fullness.
  • some depth sensors 1030 may output the minimum range distance for any object detected within the minimum range, which would result in an over-estimation, as the system may assume packages are loaded in region 1010. And certainly other example scenarios could be listed here as well.
  • detecting missing-data occlusions includes carrying out sub-process 1100 as shown in FIG. 11.
  • sub-process 1100 includes the steps of receiving a projected 2D image of grid elements at step 1101 and creating a binarization map at step 1102, performing at least one morphological opening at step 1104, performing edge detection of the at least one morphological opening at step 1106, and determining occlusion contours based on the detected edges at step 1108.
  • the binarization map delineates between (i) grid elements for which the respective depth value is valid and (ii) grid elements for which the respective depth value is not valid (i.e., a map of valid data points and invalid (e.g., missing) data points).
  • performing the morphological opening in step 1104 includes identifying a cluster of grid elements in the binarization map for which the respective depth value is invalid.
  • the identified cluster of grid elements may need to exceed a predetermined occlusion-size threshold of grid elements to be determined to be a morphological opening, and thus a (suspected) close occluding object.
  • the edge detection performed at step 1106 may be Canny edge detection.
  • performing the edge detection at step 1106 may include determining the set of grid elements in the identified cluster that define the edges of the 2D grid image after morphological opening is performed. In some embodiments, this is done on a grid element-by-grid element basis.
  • the grid elements are single pixels in the point cloud. In some embodiments, the grid elements are groups of pixels, and may be averaged together (or otherwise characterized using a single number, as described above).
  • determining the occlusion contour at step 1108 includes forming an occlusion mask (including occlusion location, contour length, and a mask image, in some embodiments) based on grid elements that have been identified as edges in the previous step.
  • the occlusion contours are based on contour length and aspect ratio.
  • the occlusion location, contour length, and mask image may be output for occlusion correction (e.g., in step 906 of FIG. 9).
  • far occluding objects The second type of occluding objects discussed herein are far occluding objects (see, e.g., FIG. 10, object 1015).
  • far occluding objects may include either or both of two different types of far occlusions: moving occlusions and discontinuous occlusions.
  • Far occlusions may be caused by occluding objects that are further away from depth sensor 1030 (i.e., closer to the loaded packages in the container) as compared with occluding objects that present as what are characterized in this description as being near occlusions.
  • Far occlusions may result in a calculated fullness that is overestimated.
  • the method for calculating the fullness assumes the shipping container has been loaded from the back to front.
  • a discontinuous occlusion e.g., a loader or a package that has not been packed yet
  • the system may assume that there are packages in the region 1020 behind the occluding object 1015, when in reality some of the space behind the object 1015 may be unoccupied.
  • FIG. 12 illustrates a sub-process of detecting discontinuous occlusions (generally at 1205) and moving occlusions (generally at 1210).
  • identifying a discontinuous occlusion includes identifying, at step 1207, a cluster of grid elements from a single frame of a 3D point cloud, where the cluster of grid elements has depth values that are more than a threshold difference less than a depth value of a loaded-portion boundary of the shipping container (i.e., the depth values for the loaded packages).
  • the cluster of discontinuous occluding points is identified using clustering techniques that are commonly know to those of ordinary skill in the art.
  • identifying discontinuous occlusions may include finding clusters with location and geometric constraints such as cluster width, length, and aspect ratio in step 1209, and may further involve confirming that the identified cluster of grid elements exceeds a predetermined occlusion-size threshold.
  • identifying the discontinuous occlusion includes performing edge detection on the cluster of grid elements.
  • identifying the discontinuous occlusion includes performing contour identification on the cluster of grid elements.
  • the grid elements are single pixels, while in other embodiments the grid elements are groups of pixels.
  • objects e.g., a loader
  • far occlusions objects that are close to the loaded packages
  • temporal analysis may be used to detect moving occlusions.
  • the transient nature of a given object may be used in identifying that object as being a moving occlusion. In some embodiments, this transient nature may be perceived as depth values changing too much between two adjacent frames for a given grid element, which may indicate movement instead of a permanently loaded package in the corresponding location.
  • the temporal analysis step includes identifying that the depth value associated with a single grid element decreases with respect to previous frames and then increases in succeeding frames in less than a threshold amount of time across multiple depth frames, consistent with what would occur if a transient object passed through the field of view of the 3D depth sensor.
  • FIG. 13 depicts an example of temporal analysis, in accordance with some embodiments.
  • FIG. 13 depicts a graph of depth values of an example individual grid element in five temporally proximate depth frames, depicted as corresponding with time intervals tl-t5.
  • each time interval may be 1/10 of a second.
  • the depth value at t3 has exceeded an example threshold depth change between at least one of the t2 depth value and the t4 depth value, and therefore the grid element in depth frame t3 may be determined to be part of a moving occlusion.
  • detecting a moving occlusion includes analyzing multiple grid elements in proximity to the detected moving occlusion grid element in order to detect a full far occluding object.
  • the upper limit of the fullness level may have a predetermined change threshold between adjacent depth frames, i.e., if a change of the estimated fullness level exceeds a predetermined limit, it may indicate the presence of a loader, for example. In other words, if a loader is relatively near the depth sensor in a region of the shipping container that hasn't been loaded yet (but not so near to the depth sensor as to cause missing data), there may be a large spike in shipping-container fullness estimation if that transient occlusion were not detected and corrected for.
  • the one or more identified occlusions corresponds to an occlusion set of the grid elements in the depth frame
  • correcting the one or more occlusions in the depth frame using one or more temporally proximate depth frames includes overwriting the occlusion set in the depth frame with data from corresponding non-occluded grid elements from one or more of the temporally proximate depth frames.
  • the non-occluded grid elements of the most adjacent depth frame may be used to fill in the occlusion set of grid elements in the current occluded depth frames.
  • FIGs. 14A and 14B depict examples of shipping-container- fullness estimation without occlusion correction and with occlusion correction, respectively.
  • the x-axis of FIGs. 14A and 14B represents time, while the y-axis represents the current shipping-container fullness estimation.
  • FIG. 14A includes results of uncorrected-for occlusions (such as between times -150-225). Methods described herein detect and correct for these occlusions, and a more accurate shipping-container fullness estimation over time is achieved, as shown by the smoothly increasing curve of FIG. 14B.
  • This disclosure proposes solutions for these two types of occlusions respectively.
  • gaps are detected from the 3D depth data.
  • Several geometric constraints including contour length, aspect ratio of the gaps are used to identify true occlusions.
  • clustering and temporal analysis may be used to identify such occlusions.
  • Container-fullness level needs to be estimated reliably even when occlusions are present.
  • the 3D depth data are corrected based on temporal analysis of multiple loading frames after the occlusions are identified. Specifically, each frame is compared with its adjacent frames, and the occluded areas are "filled" with data from corresponding non-occluded areas from adjacent frames. The fullness level is then estimated from the corrected data.
  • processors such as microprocessors, digital signal processors, customized processors and field programmable gate arrays (FPGAs) and unique stored program instructions (including both software and firmware) that control the one or more processors to implement, in conjunction with certain non-processor circuits, some, most, or all of the functions of the method and/or apparatus described herein.
  • processors or “processing devices”
  • FPGAs field programmable gate arrays
  • unique stored program instructions including both software and firmware
  • some or all functions could be implemented by a state machine that has no stored program instructions, or in one or more application specific integrated circuits (ASICs), in which each function or some combinations of certain of the functions are implemented as custom logic.
  • ASICs application specific integrated circuits
  • an embodiment can be implemented as a computer-readable storage medium having computer readable code stored thereon for programming a computer (e.g., comprising a processor) to perform a method as described and claimed herein.
  • Examples of such computer-readable storage mediums include, but are not limited to, a hard disk, a CD-ROM, an optical storage device, a magnetic storage device, a ROM (Read Only Memory), a PROM (Programmable Read Only Memory), an EPROM (Erasable Programmable Read Only Memory), an EEPROM (Electrically Erasable Programmable Read Only Memory) and a Flash memory.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Quality & Reliability (AREA)
  • Geometry (AREA)
  • Fluid Mechanics (AREA)
  • Length Measuring Devices By Optical Means (AREA)
  • Length Measuring Devices With Unspecified Measuring Means (AREA)
  • Measurement Of Levels Of Liquids Or Fluent Solid Materials (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Image Analysis (AREA)

Abstract

L'invention concerne un procédé et un appareil permettant de recevoir une trame de profondeur en provenance d'un capteur de profondeur orienté vers une extrémité ouverte d'un conteneur d'expédition, la trame de profondeur comprenant une pluralité d'éléments de grille ayant chacun une valeur de profondeur respective, à identifier une ou plusieurs occlusions dans la trame de profondeur, à corriger ladite occlusion dans la trame de profondeur à l'aide d'une ou plusieurs trames de profondeur proches temporellement, et à délivrer la trame de profondeur corrigée pour l'estimation de remplissage.
PCT/US2016/061279 2015-11-18 2016-11-10 Procédés et systèmes d'estimation de remplissage de conteneur WO2017087244A1 (fr)

Priority Applications (6)

Application Number Priority Date Filing Date Title
PL430920A PL239620B1 (pl) 2015-11-18 2016-11-10 Sposoby i układ do szacowania napełnienia pojemnika
CA3005452A CA3005452C (fr) 2015-11-18 2016-11-10 Procedes et systemes d'estimation de remplissage de conteneur
PL426752A PL426752A1 (pl) 2015-11-18 2016-11-10 Sposoby i układy do szacowania napełnienia pojemnika
GB1807994.7A GB2558507B (en) 2015-11-18 2016-11-10 Methods and systems for container fullness estimation
MX2018006105A MX2018006105A (es) 2015-11-18 2016-11-10 Metodos y sistemas para estimacion de llenado de recipientes.
DE112016005287.1T DE112016005287T5 (de) 2015-11-18 2016-11-10 Verfahren und systeme zur behälter-vollheitsschätzung

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
US14/944,860 US9940730B2 (en) 2015-11-18 2015-11-18 Methods and systems for automatic fullness estimation of containers
US14/944,860 2015-11-18
US14/978,367 US10713610B2 (en) 2015-12-22 2015-12-22 Methods and systems for occlusion detection and data correction for container-fullness estimation
US14/978,367 2015-12-22

Publications (1)

Publication Number Publication Date
WO2017087244A1 true WO2017087244A1 (fr) 2017-05-26

Family

ID=57389569

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2016/061279 WO2017087244A1 (fr) 2015-11-18 2016-11-10 Procédés et systèmes d'estimation de remplissage de conteneur

Country Status (6)

Country Link
CA (1) CA3005452C (fr)
DE (1) DE112016005287T5 (fr)
GB (1) GB2558507B (fr)
MX (1) MX2018006105A (fr)
PL (2) PL239620B1 (fr)
WO (1) WO2017087244A1 (fr)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2019125614A1 (fr) * 2017-12-22 2019-06-27 Symbol Technologies, Llc Calcul de densité de paroi de boîtier dans un chargement de remorque commerciale
FR3076897A1 (fr) * 2017-12-29 2019-07-19 Symbol Technologies Llc Systeme et procedes de generation de motif d'eclairage pour des systemes de temps de vol 3d
BE1025931B1 (nl) * 2017-12-22 2020-04-01 Symbol Technologies Llc Systemen en werkwijzen voor het bepalen van het gevuld zijn van een commerciële trailer
CN111442814A (zh) * 2020-05-13 2020-07-24 武汉市异方体科技有限公司 一种非接触式测量异形物体排水体积的装置
CN112874927A (zh) * 2021-02-03 2021-06-01 四川物联亿达科技有限公司 物流包装箱箱型推荐方法
CN113657191A (zh) * 2021-07-26 2021-11-16 浙江大华技术股份有限公司 堆积物识别方法、装置和电子装置
DE102021127789A1 (de) 2021-10-26 2023-04-27 Zf Cv Systems Global Gmbh Verfahren zur Erfassung von Ladegut
EP4212821A4 (fr) * 2020-09-08 2024-02-28 Panasonic Intellectual Property Management Co., Ltd. Procédé de mesure de taux de remplissage, dispositif de traitement d'informations et programme

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11010915B2 (en) 2019-07-11 2021-05-18 Zebra Technologies Corporation Three-dimensional (3D) depth imaging systems and methods for dynamic container auto-configuration

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060239558A1 (en) * 2005-02-08 2006-10-26 Canesta, Inc. Method and system to segment depth images and to detect shapes in three-dimensionally acquired data
EP2302564A1 (fr) * 2009-09-23 2011-03-30 Iee International Electronics & Engineering S.A. Génération dynamique d'images de référence en temps réel pour système d'imagerie à plage
US20140055560A1 (en) * 2012-08-24 2014-02-27 Microsoft Corporation Depth Data Processing and Compression
US20140372182A1 (en) * 2013-06-17 2014-12-18 Motorola Solutions, Inc. Real-time trailer utilization measurement

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060239558A1 (en) * 2005-02-08 2006-10-26 Canesta, Inc. Method and system to segment depth images and to detect shapes in three-dimensionally acquired data
EP2302564A1 (fr) * 2009-09-23 2011-03-30 Iee International Electronics & Engineering S.A. Génération dynamique d'images de référence en temps réel pour système d'imagerie à plage
US20140055560A1 (en) * 2012-08-24 2014-02-27 Microsoft Corporation Depth Data Processing and Compression
US20140372182A1 (en) * 2013-06-17 2014-12-18 Motorola Solutions, Inc. Real-time trailer utilization measurement

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2019125614A1 (fr) * 2017-12-22 2019-06-27 Symbol Technologies, Llc Calcul de densité de paroi de boîtier dans un chargement de remorque commerciale
BE1025931B1 (nl) * 2017-12-22 2020-04-01 Symbol Technologies Llc Systemen en werkwijzen voor het bepalen van het gevuld zijn van een commerciële trailer
US10628772B2 (en) 2017-12-22 2020-04-21 Symbol Technologies, Llc Computing package wall density in commercial trailer loading
GB2583245A (en) * 2017-12-22 2020-10-21 Symbol Technologies Llc Computing package wall density in commercial trailer loading
AU2018391957B2 (en) * 2017-12-22 2021-07-01 Symbol Technologies, Llc Computing package wall density in commercial trailer loading
GB2583245B (en) * 2017-12-22 2022-12-28 Symbol Technologies Llc Computing package wall density in commercial trailer loading
FR3076897A1 (fr) * 2017-12-29 2019-07-19 Symbol Technologies Llc Systeme et procedes de generation de motif d'eclairage pour des systemes de temps de vol 3d
CN111442814A (zh) * 2020-05-13 2020-07-24 武汉市异方体科技有限公司 一种非接触式测量异形物体排水体积的装置
EP4212821A4 (fr) * 2020-09-08 2024-02-28 Panasonic Intellectual Property Management Co., Ltd. Procédé de mesure de taux de remplissage, dispositif de traitement d'informations et programme
CN112874927A (zh) * 2021-02-03 2021-06-01 四川物联亿达科技有限公司 物流包装箱箱型推荐方法
CN113657191A (zh) * 2021-07-26 2021-11-16 浙江大华技术股份有限公司 堆积物识别方法、装置和电子装置
DE102021127789A1 (de) 2021-10-26 2023-04-27 Zf Cv Systems Global Gmbh Verfahren zur Erfassung von Ladegut

Also Published As

Publication number Publication date
CA3005452C (fr) 2020-07-14
PL430920A1 (pl) 2019-11-04
GB2558507B (en) 2022-04-20
GB2558507A (en) 2018-07-11
PL239620B1 (pl) 2021-12-20
DE112016005287T5 (de) 2018-08-02
PL426752A1 (pl) 2019-03-25
MX2018006105A (es) 2018-08-24
CA3005452A1 (fr) 2017-05-26
GB201807994D0 (en) 2018-07-04

Similar Documents

Publication Publication Date Title
US10713610B2 (en) Methods and systems for occlusion detection and data correction for container-fullness estimation
CA3005452C (fr) Procedes et systemes d'estimation de remplissage de conteneur
US10229509B2 (en) Methods and systems for automatic fullness estimation of containers
AU2018261257B2 (en) Method and apparatus for object status detection
US10663590B2 (en) Device and method for merging lidar data
CA2967025C (fr) Procede non parametrique d'estimation des dimensions d'objets a formes arbitraires et systeme pour l'estimation des dimensions d'objets a formes arbitraires
US8964189B2 (en) Three-dimensional measurement apparatus, method for three-dimensional measurement, and computer program
US9552514B2 (en) Moving object detection method and system
CA3152570C (fr) Systemes et procedes de guidage automatique d'installation par camera (cig)
US20180182088A1 (en) Automatic Detection, Counting, and Measurement of Lumber Boards Using a Handheld Device
Fan et al. Shelf detection via vanishing point and radial projection
WO2016158438A1 (fr) Appareil, procédé et programme de traitement d'inspection
US9639763B2 (en) Image target detecting apparatus and method
WO2013095822A1 (fr) Système de détection de ligne dynamique pour processeurs à mémoire interne limitée
CA3119342C (fr) Detection d'objet a profondeur mixte
BE1028059B1 (nl) Vervoerder-segmentatie voor datavastlegsysteem
CN110766646A (zh) 展示架遮挡检测方法、装置以及存储介质
US20240177260A1 (en) System and method for three-dimensional scan of moving objects longer than the field of view
US20230044858A1 (en) Container Fullness Tracking Method and Systems
JP6055307B2 (ja) 対応点探索装置およびカメラ姿勢推定装置ならびにこれらのプログラム
WO2019212882A1 (fr) Procédé et appareil de dimensionnement de fret à l'aide d'un rideau laser
US20210133470A1 (en) Hierarchical systems and methods for automatic container type recognition from images
EP4177848A1 (fr) Procédé et système de détermination de position d'objet contextuel
CN116645408A (zh) 一种基于点云数据的体积测量方法及装置
GB2601940A (en) Methods, systems and apparatus for segmenting and dimensioning objects

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16798918

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 3005452

Country of ref document: CA

WWE Wipo information: entry into national phase

Ref document number: MX/A/2018/006105

Country of ref document: MX

ENP Entry into the national phase

Ref document number: 201807994

Country of ref document: GB

Kind code of ref document: A

Free format text: PCT FILING DATE = 20161110

WWE Wipo information: entry into national phase

Ref document number: P.426752

Country of ref document: PL

WWE Wipo information: entry into national phase

Ref document number: 112016005287

Country of ref document: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16798918

Country of ref document: EP

Kind code of ref document: A1