Parcel number detection method and device and storage medium
A detection method and package technology, applied in the direction of measuring devices, weighing, image data processing, etc., can solve problems such as wrong weight of sorting line, irregular placement, and undue expenditure of customers, so as to reduce freight loss, The effect of high detection efficiency and precision
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0048] Taking express delivery companies as an example, express delivery companies with more advanced operating technologies are equipped with automatic sorting lines, which can sort packages according to their routing information (destination). The automatic sorting line is equipped with weighing equipment and photographing equipment, which are used to collect on-site weight and on-site images when packages are weighed.
[0049] In this embodiment, the on-site weight refers to the weight of a group of packages collected on the scene of the automatic sorting line; the on-site image refers to the image collected on the scene of the automatic sorting line; since the background is inevitably collected at the same time as the image is collected , so the live image will include both the package itself and a part of the live background where the package is located.
[0050] see figure 1 , the following is a detailed introduction to a package quantity detection method disclosed in E...
Embodiment 2
[0065] Embodiment 1 can realize the detection of the quantity of stacked packages in most cases, but there may be room for further improvement of the detection capability in cases of serious stacking. Based on this, Embodiment 2 discloses a package quantity detection method with superior detection capability.
[0066] see figure 2 , a package quantity detection method disclosed in Embodiment 2, which includes steps S10 to S40 in Embodiment 1, the improvement is that steps S50-S60 are also included after step S40:
[0067] S50: If it is recognized that the number of packages is a single piece according to the bounding box information, crop the on-site image based on the bounding box information to obtain a package image.
[0068] With regard to the identification result of the aforementioned step S40, if the number of identified packages is a single piece, it is considered that there may be cases where multiple pieces (packages with serious stacking) are misidentified as a si...
Embodiment 3
[0075] It can be understood that, in Embodiment 1 and Embodiment 2, the target detection model is by default a model that has been trained and matured. And for the training of the target detection model, there are already mature solutions in the prior art. However, Embodiment 3 further discloses a training scheme more suitable for the target detection model involved in the present invention.
[0076] A package quantity detection method disclosed in Embodiment 3 includes any solution of Embodiment 1 or Embodiment 2, that is, on the basis of Embodiment 1 or Embodiment 2, only the training method of the target detection model is improved.
[0077] see image 3 , in this embodiment, the target detection model is used to perform target prediction in step S30 after being trained through the following steps L11-L14:
[0078] L11: Annotate the bounding box information of the parcel in the live image for training.
[0079] Specifically, before the target detection model is officiall...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


