Design of portable bacterial vaginosis detection system

Abstract

Bacterial vaginosis (BV) is one of the most common diseases for reproductive age women. BV is usually diagnosed by observing the color changes and ranges of various indicators in the sample, and then obtain further information. So, the identification of sample color is very important. In terms of the traditional testing method, the acquired sample image and the calculated RGB value are used as the basis for judging the condition. Since the RGB value fluctuates greatly due to the external interference such as light source, we propose a new testing method by using the H value in HSV (color model) as the basis for identifying the color of the sample, and develop a novel BV detector based on the Raspberry Pi hardware system and digital image processing algorithm, which realizes the extraction of the color information of the sample. The BV detector mainly encapsulates the Raspberry Pi, camera, LED light source and sample platform. They are all packaged in a homemade box with the small volume of 100 x 100 x 100 mm3, which makes the instrument portable. The BV detection application is designed based on Python programming language combined with OpenCV image processing database to realize Hough circle detection and Hue value of HSV color model for color detection. The experiment results show that the BV instrument has high sensitivity, high accuracy and good stability. It is a good early BV diagnostic instrument and has important application prospect.