A method for detecting the falling of vending machine goods based on dual cameras

Abstract

The invention discloses a method for detecting goods falling of an automatic vending machine based on dual cameras. The detection steps are as follows: (1) When the equipment detects that the user's payment is completed, the pan-tilt rises to the floor where the goods are located and stops. Side camera capture; (2) Start the local image processing module of the machine tool to process the captured image, first cut the captured image according to the pre-measured position of the black and white checkerboard on the side plate of the pan / tilt, remove irrelevant areas, and then grayscale the cropped image , and finally, use the sobel edge detection operator to perform edge detection, generate an edge image, and save the edge image as a background image; (3) After the background is photographed, start the conveyor belt of the cargo lane, and start the cameras on both sides of the PTZ to shoot the PTZ (4) Stop the conveyor belt and close the camera; the present invention can detect smaller goods, flat goods, detect transparent objects, and effectively avoid the influence of ambient light on goods detection.

Description

Technical field[0001]The present invention relates to a cargo drop detection method of an automatic vending machine, in particular to a dual-camera-based automatic vending machine cargo drop detection method.Background technique[0002]The vending machine has the characteristics of simple cloth, low venue requirements, saving labor costs, is gradually become the mainstream of the new retail industry.[0003]The shipping of the automatic soldier is controlled by the machine, and whether the goods will be successfully dropped will directly affect the interests of the merchant and the user experience. Therefore, the drop detection of goods is a very important issue to be solved by the automatic galate.[0004]The traditional cargo drop detection method is generally divided into three types: the first is the infrared raster method, and the method is applied to the transmitting and receiving device of the infrared rays at both ends of the cargo cloud, forming infrared gratings, when there is g...

AI Technical Summary

Problems solved by technology

[0004] Traditional cargo drop detection methods are generally divided into three types: the first is the infrared grating method, which installs infrared emitting and receiving devices at both ends of the cargo platform to form an infrared grating. When placed on the platform, the infrared rays will be blocked, and the system will determine whether the goods have fallen according to the degree of blocking. The disadvantages of this method are: (1) there is a gap between the infrared sensors, which cannot detect slender objects; (2) the infrared sensors can only Detect objects on the grating plane. If the goods are flat (such as chewing gum) and are under the detection plane, they cannot be detected; (3) For transparent goods (such as glass cups), they cannot effectively block infrared rays and are difficult to detect

[0005] The second is the weighing method of the pan-tilt. In this method, a weighing device is installed on the pan-tilt to determine whether there is any cargo falling by detecting the weight change of the pan-tilt. The problem with this method is that the weighing sensitivity is difficult to control. If the sensitivity is too high, it is easy to cause misjudgment; if the sensitivity is too low, it will not be able to detect light weight goods

[0006] The third is the picture comparison method. This method first takes an empty pan-tilt in advance to form a standard picture, and then captures the pan-tilt during the falling of the goods, and finally compares the difference between the captured picture and the standard picture, and judges whether the goods have fallen. , the problems of this method are: (1) the patent uses a preset standard picture, and for vending machines with transparent panels, due to the influence of external light, the standard pictures taken in different time periods and in different weathers are often very different , if the lighting environment of the picture taken by the whereabouts of the goods is inconsistent with the preset picture, it is easy to cause misjudgment; (2) the patent proposes to install the camera on the top of the vending machine. It is far away, so the objects captured are small, and it is difficult to distinguish unless a high-definition camera is used; (3) The vending machine uses a preset time to control the delivery of the conveyor belt (spring); Especially after multiple shipments, the accumulation of empty trips will increase, and it is difficult to determine the shipping time; if the preset time is too long, it will easily cause over-shipments and cause losses to the merchant; if the preset time is too short, the will cause the shipment to fail

Images