Image recognition method, image collection and recognition method and commodity recognition method

Abstract

The invention provides an image recognition method, an image collection and recognition method and a commodity recognition method. The method comprises the steps: determining a first image area corresponding to an object in a first image; determining a second image area corresponding to the object in a second image based on the first image area; identifying an object in the second image area, wherein the first image and the second image are obtained by collecting the position of the object according to different angles. According to the image recognition method, the image processing process issimplified by reducing the image recognition area, so that the image recognition process is more efficient.

Description

Technical field [0001] This application relates to the field of machine vision technology, and in particular to an image recognition method. This application also relates to an image recognition device, an image acquisition and recognition method and device, a commodity recognition method and device, three computing devices, and three computer-readable storage media. Background technique [0002] With the continuous development of the concept of unmanned retail, the market for unmanned retail is growing. The problem of unmanned shelves is serious and it is difficult to promote and replicate. Therefore, unmanned containers have become a more popular option. [0003] At present, the identification technologies used by unmanned containers mainly include: RFID (Radio Frequency Identification) technology, gravity sensing technology and machine vision technology. The RFID identification scheme is to affix an RFID tag to the back of each item, and the item is purchased when it is taken ...

AI Technical Summary

Problems solved by technology

The RFID identification scheme is to put an RFID tag on the back of each item, and the item will be purchased when it is taken out of the monitoring area. The advantage is that the item is traceable and the identification result is accurate, but the operating cost is high and it is not suitable for large-scale promotion; the cost of gravity sensing technology Low, but it cannot accurately identify items of the same quality; the combination of monocular camera and gravity sensor is to use the gravity sensor to trigger the monocular camera to take pictures to identify the item, but only for static identification, the contact method between the item and the gravity sensor And the accuracy of the gravity sensor is not enough to easily affect the recognition results; therefore, pure machine vision technology becomes the preferred solution

[0004] The use of machine vision technology requires a high-threshold visual recognition algorithm. The current monocular visual recognition technology is to perform image recognition in the entire image plane. The recognition range is too wide, and the object extraction is not accurate and fast enough, resulting in a large amount of calculation when performing image recognition. Larger; monocular vision refers to the reconstruction of three-dimensional space through images captured by a single camera. For a monocular visual recognition system, when an object is blocked in the perspective of a monocular camera, the recognition fails

Images