A rapid identification method of slightly rotten fruit based on odor spectrum imaging technology

Abstract

The invention discloses a slightly decayed fruit quick identification method based on smell spectral imaging technology, which includes the steps of (1) according to the smell characters of fruits, screening m porphyrin or metal porphyrin color-sensitive materials to produce a sensor array, determining a balance time responding to the smell of the fruits, and obtaining sensor arrays before and after a reaction; (2) by means of a hyperspectral imaging system, collecting a hyperspectral image of the sensor array, performing pivot element dimension reduction, and preferably selecting n characteristic wave lengths; (3) designing a multispectral system in a multiple wave filter sheet type, obtaining n characteristic wave length images, and calculating the difference value image before and after the equilibrium reaction; and (4) extracting m*n characteristic variables on m color-sensitive points from the difference value image, constructing a determination model through mode identification, and performing quick detection to decayed fruit. Compared with a wide frequency spectrum color image obtained through CCD or a scanning instrument in the prior art, the multispectral image allows the spectral difference of unique smell molecules after effect with the color-sensitive materials, thereby increasing precision and sensitivity of detection.

Description

technical field [0001] The invention independently designs n characteristic single-wavelength filter multi-spectral systems, and uses innovative odor spectrum imaging technology to realize the detection of fruit rot. Background technique [0002] Fruits of different ripeness have more obvious odor characteristics, which are generally judged by human olfactory senses. However, the sensory evaluation is easily affected by factors such as the experience, mental state, gender, physical fitness and even the geographical environment of the reviewers, resulting in deviations in the review results. However, traditional physical and chemical detection methods are time-consuming and laborious, and it is difficult to achieve real-time detection. [0003] Most of the electronic nose systems currently used in the evaluation of fruit maturity are based on metal oxide sensor arrays, but there are certain defects in the actual use of this technology. First of all, the metal oxide sensor a...

AI Technical Summary

Problems solved by technology

However, because ordinary CCDs or scanners finally get a color RGB image, it is composed of images of three broadband bands (R, G, and B), and its spectrum is wide, covering up gas molecules and color-sensitive materials. Spectral changes at specific wavelengths after the reaction, so the accuracy of traditional electronic nose technology in food odor detection needs to be further improved, and its application is limited

Images