Foreign-matter detection algorithm of liquid in transparent glass bottle

Abstract

The invention provides a foreign-matter detection algorithm of liquid in a transparent glass bottle. Firstly, guided filtering enhancement is carried out on video images; then ''Hough'' transformationis used to detect an outline of the transparent glass bottle, and an algorithm detection region is reduced to a liquid detection area; a background model is built for each pixel of the liquid detection region; similarity of a current pixel point and the background model is calculated, if the pixel point is similar to the background model, the same is classified as a part of a background, and otherwise, the same is a part of a foreground; and the area of a connected region of the foreground is calculated, if the area of the connected region is less than a set threshold value, the same is classified as the background, and otherwise, the same is the foreground. The method can detect foreign matters in non-transparent liquid.

Description

technical field [0001] The invention relates to an image processing technology, in particular to a foreign matter detection algorithm for liquid in a transparent glass bottle. Background technique [0002] In the intelligent detection equipment for non-transparent liquid products in transparent glass bottles, the core detection technology is product defect recognition and detection technology based on machine vision, and better image enhancement processing is the basis for subsequent detection of foreign objects in liquids. [0003] At present, there are four image enhancement techniques with better enhancement effects: histogram equalization, dark channel method, Retinex_MSR and guided filtering method. [0004] (1) Histogram equalization is a histogram correction method based on cumulative distribution function transformation, which can generate an image with uniform gray level distribution probability, and the image after histogram equalization has a large amount of infor...

AI Technical Summary

Problems solved by technology

[0005] (2) The dark channel prior algorithm has a good effect on single image dehazing processing, but the processing time is long and requires a lot of storage resources and computing resources

Therefore, the disadvantage of this algorithm is that the real-time performance is poor, and the image processed by the dark channel prior algorithm will be darker than the original image, and the processed image needs to be exposed, which again increases the complexity of the algorithm.

When dealing with color images, the MSR algorithm often suffers from insufficient color saturation and halos.

Images