CN-121981142-A - DWS cargo information binding method and system based on visual target tracking

CN121981142ACN 121981142 ACN121981142 ACN 121981142ACN-121981142-A

Abstract

The invention discloses a DWS cargo information binding method and system based on visual target tracking, which are characterized in that cargo image data and depth data of a full operation area of a DWS system are acquired through a plurality of image acquisition devices, and a panoramic image sequence covering the full area is generated through splicing and fusion; the method comprises the steps of carrying out real-time detection and tracking on target cargoes in a panoramic image sequence based on a target detection model, obtaining three-dimensional physical coordinates of the cargoes under a unified physical coordinate system, completing association binding of bar code information and corresponding cargoes through matching with a reference coordinate, and correspondingly executing binding or unbinding of bound information of the cargoes and RFID labels of the transportation carriers when the cargoes enter/leave a transportation carrier loading/unloading area based on a visual tracking result. The method effectively solves the problems of tracking interruption and target confusion caused by goods accumulation, shielding and trans-regional movement, ensures the continuity and stability of goods tracking, and remarkably improves the accuracy and scene suitability of goods identity information binding.

Inventors

SONG HONGQING
SONG HAO
CHEN YU
TANG GUOHAO
HE YIN
WU JIE
ZHOU XINYUE
HE CHENG
TANG QIUWEN
CHENG PENGZHI
LUO JIA
ZHANG YIWEN
Bai Marongzhong
DU MINGQIAN
LI SILIN
LI TAO
FENG WENFEI
ZHU KEYAN

Assignees

民航成都物流技术有限公司

Dates

Publication Date: 20260505
Application Date: 20260408

Claims (10)

1. A DWS cargo information binding method based on visual target tracking, applied to a DWS system for logistics freight transportation, the DWS system including a code reading camera for acquiring cargo bar code information, the method comprising the steps of: Acquiring cargo image data and depth data in an operation area through a plurality of image acquisition devices deployed in the operation area of a DWS system; splicing and fusing the acquired cargo image data and depth data to generate a panoramic image sequence covering the operation area; real-time detection and tracking are carried out on target cargoes in the panoramic image sequence based on a target detection model, and three-dimensional physical coordinates of each target cargo are calculated by using pose information of the code reading camera and the image acquisition device and the depth data according to a pre-established unified physical coordinate system; According to the unified physical coordinate system, mapping the pixel coordinates corresponding to the bar code information acquired by the code reading camera into bar code physical coordinates, and when the bar code physical coordinates are matched with the three-dimensional physical coordinates of the target goods, associating and binding the bar code information with the image data of the corresponding target goods; When the tracked target goods are detected to enter a transport carrier loading and unloading area in the operation area, the association binding of the bound information of the target goods and the transport carrier RFID tag is executed, and when the tracked target goods are detected to leave the transport carrier loading and unloading area, the unbinding of the bound information of the target goods and the transport carrier RFID tag is executed.
2. The method for binding DWS cargo information based on visual target tracking according to claim 1, wherein the image capturing device includes RGBD cameras, and a plurality of the RGBD cameras are distributed on two sides of the cargo conveyor belt, a platform edge, and a transport vehicle loading and unloading area.
3. The DWS cargo information binding method according to claim 1, wherein the performing stitching and fusion on the collected cargo image data and depth data to generate a panoramic image sequence covering the working area includes: Binding a global unified time stamp to the image frames acquired by the image acquisition device and performing time sequence calibration; And carrying out multi-scale feature extraction on the image data subjected to time sequence calibration by adopting a SWIN-transducer network, and completing feature fusion and video stitching of multi-view images through a self-attention mechanism to generate a panoramic feature map so as to obtain a panoramic image sequence covering the operation area.
4. A DWS cargo information binding method according to claim 3, wherein the target detection model uses YOLOv a depth learning model, and the YOLOv a depth learning model uses the collected depth data to detect and locate target cargos in a panoramic image sequence, so as to obtain a moving track of each target cargos in continuous image frames.
5. The DWS cargo information binding method according to claim 1, wherein the calculating three-dimensional physical coordinates of each target cargo according to a pre-established unified physical coordinate system includes: And based on the internal parameters and the external parameters obtained by calibration and depth data obtained by the image acquisition device, establishing a mapping rule of pixel coordinates of each imaging device and a unified physical coordinate system, and realizing conversion from the pixel coordinates to three-dimensional physical coordinates.
6. The DWS cargo information binding method based on visual target tracking according to claim 1, further comprising a barcode abnormality processing step of: for target goods which do not successfully acquire bar code information, marking and prompting by a red square frame on an interactive display interface, and guiding staff to complete manual repair and bar code binding; and marking the target goods with the completed bar code information association binding by a green square frame on the interactive display interface.
7. The DWS cargo information binding method based on visual target tracking according to claim 1, wherein the condition for matching the barcode physical coordinates with the three-dimensional physical coordinates of the target cargo is: Calculating the space distance between the bar code physical coordinates and the three-dimensional physical coordinates of the target goods identified by the tracking based on the three-dimensional physical coordinates of the target goods identified by the current tracking; And when the spatial distance is smaller than a preset threshold value, judging that the matching is successful.
8. The method for binding DWS cargo information based on visual object tracking according to claim 1, wherein the performing of the associated binding of the bound information of the object cargo to a carrier RFID tag when the tracked object cargo is detected to enter a carrier handling area within the work area, and the performing of the unbinding of the bound information of the object cargo to a carrier RFID tag when the tracked object cargo is detected to leave the carrier handling area, comprises: When the tracking and identifying target goods are detected to enter the transport carrier loading and unloading area of the operation area, triggering RFID identification operation to acquire RFID label information of a corresponding transport carrier, associating and binding the bound information of the target goods with the RFID label information of the corresponding transport carrier, synchronously marking the target goods to be in a loaded state, when the tracking and identifying target goods are detected to leave the transport carrier loading and unloading area of the operation area, triggering unbinding operation to release the association relation between the bound information of the target goods and the RFID label information of the corresponding transport carrier, and synchronously clearing the loaded state mark of the target goods, wherein the RFID label information is automatically scanned and identified by an RFID reader or acquired in a manual input mode.
9. A DWS cargo information binding system based on visual target tracking for performing the DWS cargo information binding method based on visual target tracking according to any of claims 1-8, the system comprising: The DWS basic module comprises a code reading camera, a dynamic weighing unit and a volume measuring unit and is used for acquiring bar code information, weight data and volume data of target goods; the multi-view vision acquisition module comprises a plurality of RGBD cameras distributed on two sides of the cargo conveyor belt, the edges of the platform and the loading and unloading areas of the transport vehicles, and is used for acquiring cargo image data and depth data in the operation area; the edge computing module is used for executing image splicing and fusion, target detection and tracking, coordinate mapping and information binding and loading and unloading state management operations; The RFID identification module comprises an RFID reader and an RFID tag arranged on the transport carrier, and is used for identifying the identity of the transport carrier and triggering the association binding and unbinding operation; And the interactive display module is used for displaying real-time information of the target goods, tracking pictures and receiving manual operation instructions.
10. The DWS cargo information binding system based on visual target tracking according to claim 9, wherein the interactive display module is further configured to mark, on the interactive display interface, cargo that has not successfully identified the barcode information with a red box; The edge calculation module is also used for fusing the bar code information, the three-dimensional physical coordinates, the moving track and the loaded state of the goods with the goods weight data acquired by the dynamic weighing unit and the volume data acquired by the volume measurement unit, generating a goods data packet and uploading the goods data packet to the management system in real time according to a preset communication protocol.

Description

DWS cargo information binding method and system based on visual target tracking Technical Field The invention relates to the technical field of intelligent logistics, in particular to a DWS cargo information binding method and system based on visual target tracking. Background The dynamic weighing and scanning system (DYNAMICS WEIGHING AND SCANNING, DWS) is used as core equipment of a logistics sorting and transferring link and is widely applied to places such as airports, express distribution centers and the like. At present, the DWS system mainly integrates three core modules of a code reading camera, a volume measuring device and a dynamic weighing unit, can complete automatic acquisition of basic information of goods in a conveying line sorting scene, and transmits acquired bar codes, weight and volume data to an upper management system to realize basic informationized management of the goods. With the high-speed development of the e-commerce logistics industry and the air freight industry, the requirements of the freight sorting and transferring links on the operation efficiency, the sorting accuracy and the whole-flow traceability are continuously improved. However, the existing DWS system still has the following technical bottlenecks in practical application: First, existing DWS systems typically deploy only a small number of code reading cameras and tracking cameras at critical nodes of the transmission line, which are limited in their field of view. The fixed side plates at two sides of the conveyor can shield the visual field of the camera to form a visual blind area, meanwhile, the visual range of the existing camera is limited by the installation position, and the visible range of the existing camera is not covered by about 20-30 cm from the edge of the platform, so that the cargo state at two sides of the platform and the loading and unloading area can not be effectively sensed. The existence of the dead zone causes the interruption of visual information of the goods in the links of conveying, transferring, loading and unloading and the like, and a complete goods track record cannot be formed. Secondly, the freight transportation line generally lacks an automatic distance pulling mechanism, and when a plurality of cargoes continuously enter or the volume difference of cargoes is large, the phenomenon of stacking and overlapping of the cargoes is easy to occur. In this case, the bar code recognition device positioned at the code reading port is easily shielded by the goods in front or above, and can not clearly collect the bar code image of the shielded goods, and meanwhile, the visual tracking system is difficult to maintain a stable tracking mark because of the shielding of the target, so that the tracking of the goods is interrupted and lost. Thirdly, the information binding accuracy and scene robustness of the bar code and the goods are seriously insufficient, and the problem that binding failure and false binding are missed easily occurs under complex working conditions is that the prior art can only realize independent calibration of a code reading camera and a tracking camera, and binding logic is highly dependent on an ideal working environment without shielding, blind areas and displacement. Once the complex working conditions such as visual blind areas, overlapping and shielding of goods, manual carrying and shifting and the like appear, the coordinate mapping reference of the bar codes and the goods can deviate, the accuracy of the coordinate matching of the bar codes and the goods can not be ensured, and then the problems of wrong binding, missing binding and invalid binding relation of the bar codes appear, so that the sorting accuracy of the goods is seriously influenced, and even the goods cannot be traced after being misplaced and lost. In summary, the barcode recognition rate, cargo tracking continuity and information binding accuracy of the existing DWS system in a complex freight scene are difficult to meet actual operation requirements, and a large amount of manual intervention is needed for repair and review, so that the operation efficiency is reduced, and further development of logistics automation is restricted. Disclosure of Invention Aiming at the defects existing in the prior art, the invention provides a DWS cargo information binding method and system based on visual target tracking, which solve the problems that blind areas exist in visual coverage and the bar codes are difficult to identify due to cargo accumulation shielding and the continuous tracking of the whole flow cannot be carried out on cargoes in the prior art, realize the visual coverage and the full flow cargo tracking of the whole area of a conveying line and improve the accuracy rate of the association of the bar codes and cargo images. According to an embodiment of the invention, a DWS cargo information binding method based on visual target tracking is applied to a DWS system of logistics freight tran