Device and method for single-camera repetitive visual detection of tobacco rods

Abstract

The invention relates to a single-camera repetitive visual detection device and method for tobacco rods, belonging to the technical field of tobacco electrical automatic detection. The single-camera repetitive visual detection device for tobacco rods includes a conveyor belt, a light source group, a camera, a repetitive trigger optical fiber group, Controller, graphics processor; the present invention triggers a camera by repeatedly triggering multiple optical fiber sensors in the optical fiber group to obtain multiple images of the same side or local different angles of the tobacco rod at different positions, the images and information are redundant, and the verification is passed , error correction, comparison, synthesis and other processing to obtain more accurate and more information, and the types of graphics processing algorithms are more reliable and richer. Thereby reducing unqualified tobacco rods that are not detected and entering into the finished product, improving product quality, and at the same time, greatly reducing the proportion of wrongly rejected qualified tobacco rods, and improving the accuracy and reliability of tobacco rod detection. Moreover, the cost is low, and one camera replaces multiple cameras to photograph the same side or part of the tobacco rod.

Description

technical field [0001] The invention belongs to the technical field of tobacco electrical automatic detection, and in particular relates to a single-camera repetitive visual detection device and method for tobacco rods. Background technique [0002] Existing tobacco rod visual appearance inspection devices use multiple cameras to photograph and detect multiple faces or parts of the tobacco rod, but for each face or part of the tobacco rod, one camera is used to inspect one face or part of the tobacco rod Take one shot, one image, one graphics processing and analysis to find out whether the surface or part is qualified. The information obtained from an image is too little, and when there are various interferences, errors, partial unclear images, etc. in the process of image shooting, transmission, and processing, the detection result is inaccurate and unreliable. [0003] Use a camera to shoot a surface or part of a tobacco rod, an image, and a graphic processing and analysi...

AI Technical Summary

Problems solved by technology

The information obtained from an image is too little, and when there are various interferences, errors, partial unclear images, etc. in the process of image shooting, transmission, processing, etc., the detection result is inaccurate and unreliable

[0003] Use a camera to shoot a surface or part of a tobacco rod, an image, and a graphic processing and analysis. Due to the limited images and information obtained, the available detection algorithms are also limited, and the detection results obtained in many cases are uncertain. Whether the cigarette stick is qualified, but the result must be given in the end, only qualified or unqualified

[0004] For example: at a certain angle, the local area of ​​the cigarette rod image is bright and nearly saturated, and the reflection is serious. Under the limited conditions of only a single image, it is impossible to further analyze the actual situation of the reflective area. The highlighted reflective area may be a serious quality defect. In order to provide reliable products Quality, often can only be judged as unqualified and removed

But this is an important reason for the high error rejection rate. A large number of small wrinkles and slight random deformations such as transparent wrapping paper are within the allowable range of the process standard. At a certain angle, the local highlight saturation of the tobacco rod image is serious. , Excessively high false rejection rate makes the detection device lose its use value, but reducing the false rejection rate cannot guarantee product quality

Images