Long-term stable, high accuracy, and visual detection platform for In-field analysis of nitrite in food based on colorimetric test paper and deep convolutional neural networks

[Display omitted] •In-field measurement of nitrite as carcinogen is important for food safety.•A fully integrated colorimetric detection system for nitrite is offered.•APP is combined with DCNN as visual monitoring platform.•Validity period of test paper is prolonged from 7 d to more than 30 d.•The accuracy of food classification is high as 91.33–100% Nitrite is one of the most common carcinogens in daily food. Its simple, rapid, inexpensive, and in-field measurement is important for food safety, based on the requirements of the standard from Codex Alimentarius Commission and China. Using polyacrylonitrile (PAN) and thin layer silica gel (SG), p-aminophenylcyclic acid (SA) and naphthalene ethylenediamine hydrochloride (NEH), as carriers and chromogenic agents, respectively, PAN-NSS as nitrite color sensor is proposed. After fixing and protecting of SA and NEH with layer-upon-layer PAN, the validity period of the test paper can be prolonged from 7 days to more than 30 days. The reproducibility of PAN-NSS preparation is ensured by electrospinning. Combined with PAN-NSS, deep convolutional neural network (DCNN) and APP as a visual monitoring platform, which has the functions of rapid sampling, data processing and transmission, intuitive feedback, etc., and provides a fully integrated detection system for field detection.