**1. Introduction**

Crop diseases cause huge losses to agricultural production and directly affect the economic development and national food security of many countries in the world [1]. Among them, fungal diseases can cause huge losses to the growth and yield of crops. In addition, fruits will still be damaged by fungi after being picked [2]. Fungal diseases mostly exist in the form of spores before infecting crops and fruits. In addition to causing damage to crops and fruits, fungal spores can also enter the lungs through the human respiratory tract and spread to other organs of the human body, causing various fungal diseases [3,4].Therefore, it is necessary to capture and identify megaspores quickly and accurately.

The identification and counting of spores under the traditional microscope mainly depend on naked eye observation. Due to the large number of spores captured, this method is labor-intensive, time-consuming, and inefficient. The accuracy of observation depends on the professional experience of operators, sometimes leading to large errors. Image processing methods are used to automatically detect and count spores, including image segmentation using K-means clustering algorithm, recognition based on shape factor and area, and spore contour segmentation based on concavity and contour segment merging.

**Citation:** Zhang, X.; Bian, F.; Wang, Y.; Hu, L.; Yang, N.; Mao, H. A Method for Capture and Detection of Crop Airborne Disease Spores Based on Microfluidic Chips and Micro Raman Spectroscopy. *Foods* **2022**, *11*, 3462. https://doi.org/10.3390/ foods11213462

Academic Editor: Lili He

Received: 7 October 2022 Accepted: 27 October 2022 Published: 1 November 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

The automatic detection of disease spores has good effectiveness and accuracy, but spore recognition based on image processing technology cannot contain too many features [5,6]. Using deep neural network FSNet to detect fungal spores can automatically identify and count fungal spores in microscopic images, but image and deep learning methods are still not accurate enough to identify spores with similar shape and size [7]. PCR is the gold standard method for microbial detection. It is often used for microbial identification with high accuracy. However, this method requires professionals to crack the spores under strict experimental conditions. Only through tedious processing can fungal spores be detected [8,9]. The existing microscopic image method can realize rapid detection and recognition through morphology, but it cannot accurately identify spores with similar shape and size. The PCR method has high accuracy, but the detection conditions are harsh, destructive sampling is required, and the timeliness is poor, and the cost is high. Therefore, it is urgent to develop a rapid and accurate spore detection and identification technology. Raman spectroscopy is a light scattering technology with low cost and high speed. It can reflect various vibration frequencies and related vibration levels of biological components and can be used to identify the molecular composition and structure of biological samples [10,11]. Micro Raman spectroscopy has been applied to the identification and analysis of bacteria, which can be classified and identified [12,13]. However, the content of spores in the air is low, and there are a lot of impurities, so it is difficult to detect them directly using Raman spectroscopy. Therefore, a separation and enrichment method is needed to achieve spore detection and improve detection accuracy.

In recent years, microfluidic technology has provided powerful tools for detection applications due to its portability, miniaturization, automation, multi-channel sample detection, and cost saving. Compared with traditional methods, the greatest advantage of microfluidic control is to create a controllable microenvironment, which can accurately drive and control the microfluidic flow in the microchannel, thus improving the detection sensitivity [14,15]. Among them, the impactor designed according to the principle of aerodynamics is widely used in the separation of atmospheric particles with good results [16], and it is feasible to collect fungal spores using a microfluidic chip with a virtual impactor structure. This method has high collection efficiency and low cost [17,18]. In addition, microfluidic chip composite diffraction is used to detect and identify spores. For spores with large differences in size and shape, the identification accuracy is high. Compared with microscopic observation, its detection field of vision is much expanded, and the detection efficiency and speed are greatly improved [19,20]. However, the recognition accuracy of spores with similar shape and size is not high due to morphological detection. Therefore, it is necessary to design a separation and enrichment microfluidic chip with compound Raman spectroscopy to capture and collect spores and identify them accurately.

In this study, a two-stage separation and enrichment microfluidic chip with a semi arc pretreatment structure is proposed, which can be combined with micro Raman. Using numerical analysis and experiment, the best design parameters are obtained, and the feasibility of microfluidic chip is discussed. At the same time, micro Raman was used to detect the collected spores to obtain the Raman spectrum of fungal spores. The Raman spectrum was preprocessed and modeled for analysis. The fungal spores were identified, and their fingerprints were established to provide evidence for the subsequent prevention and treatment of fungal diseases.
