Electronic devices for inventory management of goods on shelves based on image recognition

Abstract

An electronic device for inventorying shelf merchandise based on image recognition provides image suggestions for restocking. The electronic device includes an input unit, a storage unit, and a processing unit. The input unit receives an inventory request, which includes multiple input images. The processing unit executes a recognition program stored in the storage unit. The recognition program includes the following steps: the recognition program stitches the input images into a shelf composite result, which has multiple sets of products to be measured; the recognition program selects a corresponding preset product set from preset display data based on the area of ​​the product sets to be measured, and draws the selected preset product set as a target product frame; the recognition program overlays the target product frame onto the shelf composite result to generate restocking prompt information.

Claims

1. An electronic device for inventorying goods on shelves based on image recognition, comprising: One input unit receives an inventory request, which includes multiple input images; An input unit receives an inventory request, which includes multiple input images; a storage unit stores multiple sets of preset display images and a recognition program, each preset display image having multiple preset shelves, multiple preset product sets, and multiple preset product areas, each preset product area being configured with a corresponding preset product set; and a processing unit connected to the input unit and the storage unit, the processing unit executing the recognition program, the recognition program listening to the inventory request, the recognition program stitching the input images into a shelf composite result, the shelf composite result having multiple product sets to be tested; the recognition program selecting the corresponding preset product set from the preset display images according to the area of ​​the product set to be tested, and drawing the selected preset product set as a target product frame; and the recognition program generating a replenishment prompt information by overlaying the target product frame onto the shelf composite result.

2. The electronic device for inventorying shelf goods based on image recognition as described in claim 1, wherein the shelf synthesis result includes multiple target shelves, each target shelf having at least one set of goods to be tested or not having the set of goods to be tested, and there is a product gap between two adjacent sets of goods to be tested, or the target shelf that does not have the set of goods to be tested has the product gap.

3. The electronic device for inventorying shelf merchandise based on image recognition as described in claim 1, wherein the recognition program adjusts the regions of multiple target shelves in the shelf synthesis result according to a first axis of the shelf synthesis result, establishes a shelf index for each target shelf, maps each shelf index to the corresponding preset shelf, the recognition program generates a first product label vector according to the preset shelf, the recognition program generates a second product label vector according to the corresponding target shelf, the recognition program generates a similarity between the first product label vector and the second product label vector, the recognition program determines that the similarity is greater than a first threshold value, and the recognition program establishes the shelf index of the target shelf.

4. The electronic device for inventorying shelf merchandise based on image recognition as described in claim 3, wherein the recognition program scans each target shelf according to a second axis of the shelf synthesis result and obtains the set of merchandise to be tested for each target shelf, wherein the first axis is perpendicular to the second axis.

5. The electronic device for inventorying shelf merchandise based on image recognition as described in claim 4, wherein the recognition program extracts an appearance feature embedding vector from each of the input images, the recognition program calculates a feature matching score for the appearance feature embedding vectors of two adjacent input images, the recognition program determines that the similarity is greater than a first threshold value and the feature matching score is greater than a second threshold value, and the recognition program confirms and establishes the shelf index of the target shelf.

6. The electronic device for inventorying shelf merchandise based on image recognition as described in claim 3, wherein the recognition program selects the corresponding preset shelf and the preset merchandise set from the preset display data according to the shelf index, and the recognition program adjusts the size of the target merchandise frame according to a merchandise gap.

7. The electronic device for inventorying shelf merchandise based on image recognition as described in claim 6, wherein the recognition program determines a replenishment quantity to fill the gaps between merchandise based on a merchandise width of the preset merchandise set.

8. The electronic device for inventorying shelf merchandise based on image recognition as described in claim 1, comprising a display unit connected to the processing unit, the display unit displaying the shelf composite result or the replenishment reminder information.

Images