Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Transform-based logistics package separation method

A separation method and wrapping technology, applied in neural learning methods, image analysis, biological neural network models, etc., to achieve time performance advantages, reduce computational complexity, and improve segmentation processing speed.

Pending Publication Date: 2022-07-05
SOUTH CHINA UNIV OF TECH
View PDF0 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

For the task of logistics package sorting, it is improved under the two indicators of segmentation accuracy and model calculation complexity

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Transform-based logistics package separation method
  • Transform-based logistics package separation method
  • Transform-based logistics package separation method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0081] see figure 1 and figure 2 , the present invention provides a kind of logistics parcel separation method based on Transformer, comprises the following steps:

[0082] Step 1. Use the camera to collect the image of the conveyor belt logistics package in real time, and complete the selection of the area of ​​interest in the system. In this embodiment, the camera type is an industrial 3D camera with a built-in depth computing chip.

[0083] Step 2. Pass the image of size H×W×C into the improved Transformer semantic segmentation model of the package separation method, such as image 3 As shown, the original image is divided into multiple tiles with local continuity between different tiles. The size of the original input image is H×W×C, where H is the image width, W is the image height, C is the number of image channels, and when the original image is an RGB image, C=3; each image is divided into different The size of the block is 7 × 7 (pixels), and the dimension of the...

Embodiment 2

[0135] In some logistics parcel sorting systems, the status of the conveyor belt remains unchanged. In the video stream captured by the top camera, the background of the picture group composed of adjacent frames is almost unchanged, so the above steps 2 to 7 can be used. The segmentation method is replaced by a Gaussian mixture model separation algorithm (MOG) to enhance the real-time performance of package detection. This embodiment is implemented with the help of the OpenCV open source library. By adjusting the parameters to adapt to the detection of the shadow by the algorithm, the interference of the shadow area in the image is eliminated, so as to pay more attention to the actual characteristics of the package. The specific steps are as follows: 1) In Embodiment 1, the camera The captured image is subjected to median filtering to remove noise; 2) The image matrix is ​​put into the GPU to speed up the detection process; 3) The MOG algorithm is used to segment the image to o...

Embodiment 3

[0138] Since this embodiment uses an industrial 3D camera with a built-in depth computing chip, the depth information of the image can be used to perform foreground segmentation to enhance the robustness of package detection. The industrial 3D camera used in this embodiment can automatically calculate depth information and output a depth image. like Figure 9 shown, the wrapped depth image captured by the depth camera in this embodiment. The basic idea of ​​using depth information for parcel segmentation is to identify the contour by the gradient of the depth map. If the depth difference between two adjacent regions is large, the gradient at the boundary will be very large and can be extracted as a contour. This embodiment is implemented with the help of the OpenCV open source library, and the specific steps are as follows: 1) perform a morphological closing operation on the depth map to reduce the holes in the image; 2) use the Sobel operator to obtain the gradient of the de...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a logistics package separation method based on Transform, and the method comprises the following steps: sending an image into an improved Transform semantic segmentation model, dividing the received image into a plurality of image blocks, and transmitting the image blocks into a hierarchical encoder, the encoder outputs multi-level image features with different resolutions by using overlapping feature merging operation and a feedforward neural network in combination with a self-attention mechanism; carrying out feature splicing and fusion by using a lightweight decoder based on a multi-layer perceptron, and predicting package segmentation mask information of the image; performing image morphological post-processing on the mask information, extracting edge information of all the packages, acquiring the distribution condition of the current packages, performing statistics on the distribution condition of the packages on the conveyor belt, and acquiring the package in the forefront of the conveyor belt as a target package; and the target parcel information is used as updating input of a Kalman filtering target tracking link, so that single-piece sorting of the logistics parcels is realized.

Description

technical field [0001] The invention belongs to the field of logistics system parcel sorting, and relates to a Transformer-based logistics parcel separation method. Background technique [0002] With the popularization of Internet technology and the prosperity of e-commerce in my country, the logistics and warehousing industry has achieved rapid development. Due to the large population in our country and the increasing number of online shopping customers, the parcel sorting work in the logistics warehouse is increasingly heavy. With the development of artificial intelligence, especially computer vision technology, there are fewer and fewer occasions for manual sorting. It has become a trend to use intelligent algorithms to assist logistics warehouses for parcel sorting to achieve warehousing automation and intelligence. [0003] At present, the computer vision methods used for package identification and sorting are mainly deep learning semantic segmentation algorithms, whic...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06T7/215G06T7/246G06T7/277G06T7/13G06T5/00G06T5/10G06T5/30G06N3/04G06N3/08
CPCG06T7/215G06T7/246G06T7/277G06T7/13G06T5/10G06T5/30G06N3/04G06N3/08G06T2207/20036G06T2207/20056G06T2207/20081G06T2207/20084G06T5/70
Inventor 谢巍秦奕别业泉谭淏周雅静
Owner SOUTH CHINA UNIV OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com