Corn Seed Quality Detection Based on Spectroscopy and Its Imaging Technology: A Review

Abstract

The quality assurance of corn seeds is of utmost significance in all stages of production, storage, circulation, and breeding. However, the traditional detection method has some disadvantages, such as high labor intensity, strong subjectivity, low efficiency, cumbersome operation, etc. In view of this, it is of great significance to study more advanced detection methods. In this paper, the application of near-infrared spectroscopy and its imaging technology in the quality detection of corn seeds was reviewed. Firstly, the principles of these two technologies were introduced, and their components, data acquisition, and processing methods, as well as portability, were compared and discussed. Then, the application of these methods to the main quality of corn seeds (including variety and purity, vigor, internal components, mycotoxins, and other qualities such as frost damage, hardness, and maturity, etc.) was reviewed. Breakthroughs and innovations have been made in detection methods, spectral preprocessing methods and recognition algorithms. The significance of corn quality characteristics and the function of the applied algorithm were emphasized. Finally, the challenges and future research direction of spectral and its imaging technology was proposed, aiming to further enhance the accuracy, reliability, and practicability of the detection technology. With the rapid development of spectral and its imaging technology, the detection methods of corn quality are also advancing with the times. This is not just for corn, but more and more crops can be accurately detected by these technologies. It will become an important means of agricultural production inspection in the future.

1. Introduction

Corn (Zea mays L.) is one of the three major food crops globally, rich in proteins, lipids, vitamins, and various nutrients. It serves both as an economic and a feed crop and plays a key role as an ingredient in the pharmaceutical, starch, and alcohol industries. According to the food and agriculture organization (FAO) statistics of the United Nations, the global yield of corn is more than 1.16 billion tons in 2022 (https://www.fao.org/faostat/zh/#data/QC, accessed on 10 November 2024), and China’s corn production in 2023 reached 288.8 million tons (https://data.stats.gov.cn/easyquery.htm?cn=C01, accessed on 10 November 2024), accounting for 41.54% of the total production of all food crops, ranking first in the planting area and output of major food crops. In addition, the planting efficiency of corn is higher than that of other food crops. It has become the main source of income for many crop farmers. Therefore, corn holds an extremely important position in China’s agricultural production. To strengthen research on corn is crucial for promoting agricultural development, economic growth, and stability.

In the process of the production, processing, and sales of corn seeds, its quality detection is particularly important, the different varieties make its use of different scenarios, some suitable for human consumption, some suitable for animal feed, and some can be made into a lot of agricultural and sideline processing products, the internal composition of corn differences also create its use. In addition, the level of seed vitality for the seedling and growth of corn is also affected. If the seed is improperly stored, it will suffer mycotoxin infection, etc., which affects life and production and even causes harm to human health. Thus, in the process of corn production, storage, circulation, and breeding, it is a key step to accurately detect corn quality of each process. According to the International Seed Testing Association (ISTA), the indicators include physical, biological, genetic, and healthy indicators such as variety and purity, vigor and component, etc. Zhang et al. clarified the advantages and disadvantages of traditional maize seed quality testing methods [1]. Table 1 shows the main detection indicators and application scenarios of corn in each production process, and the advantages and disadvantages of traditional detection methods are explained [1,2], with a list of the relevant traditional detection methods. Table 1 shows that the current traditional corn quality detection methods include manual detection, drying method, Kjeldahl nitrogen determination, spectrophotometry, DNA molecular marker method, protein electrophoresis identification, etc., most of which are based on chemical analysis. Although the detection accuracy is high, there are obvious drawbacks, such as heavy workload, strong subjectivity, low efficiency, high cost, and harm to human health. Moreover, the detection samples cannot be recycled and used, the reagents and corn samples require sufficient reaction time, which is time-consuming and labor-intensive [1,2,3,4]. The quality of corn directly affects the health of the public and its processing characteristics, so the research on its quality detection methods is of great significance.

Table 1. The main detection indicators and application scenarios of corn in production processes.

Indicators	Application Scenario	Traditional Detection Methods	Advantages	Disadvantages
Variety and purity	To avoid losses caused by shoddy and mixed seeds.	Manual, protein electrophoresis, DNA analysis, etc.	High accuracy	Destructive, time-consuming
Vigor	To improve seed survival and yield.	Germination rate, electrical conductivity, artificial accelerated aging measurement, etc.	Simple, easy to operate	Time-consuming, destructive, large sample size
Component	To provide a basis for rational use and processing.	DNA molecular markers, Kjeldahl nitrogen determination, acid hydrolysis, thermogravimetric analysis, etc.	High accuracy	Time-consuming, destructive
Moisture	To provide a basis for storage and processing	Drying, microwave heating, resistance method, etc.	High accuracy	Time-consuming, destructive
Mycotoxins	To prevent harm to humans and animals.	HPLC, LC-MS/MS method, etc.	High accuracy	Expensive, Complicated operation, destructive
Freezing damage	To provide a basis for seed selection and use.	Observation, germination, electrical conductivity method, chemical analysis	Simple, easy to operate	Destructive, time-consuming
Maturity	To determine the quality and yield potential	Germination determination, nuclear magnetic resonance method	Simple, easy to operate	Time-consuming, destructive
Hardness	To provide a basis for processing and packaging.	Hardness meter measurement, grinding method	Easy to operate	Destructive, time-consuming

Existing non-destructive corn quality detection technologies [5] include electronic nose and electronic tongue detection methods based on chemical properties, dielectric property detection methods based on electrical properties, spectral technology based on optical properties, etc. Since optical property detection technologies have high detection speed and accuracy, some research based on near-infrared spectroscopy (NIRS) and hyperspectral imaging (HSI) have been reported before, according to the characteristics of non-destructive and rapid spectral imaging technology, as well as the wide application in agriculture and food detection. Therefore, to help readers to understand the field of corn quality detection, this study collated and summarized the literature of NIRS and HSI technology in nondestructive testing of corn quality. In this review, the principles of these two technologies (NIRS and HSI) were first introduced, and the difference between them were compared. Then, the development and application of different indicators such as corn variety and purity, vitality, internal composition, mycotoxins infection, and other indicators are introduced. Finally, the challenges and future research directions in corn quality detection are discussed. The schematic diagram of this study is shown in Figure 1.

Figure 1. The schematic diagram of this study.

2. Introduction of NIRS and HSI Technologies

2.1. Basic Principles and Characteristics of NIRS

Near-infrared spectroscopy (NIRS) ranges from 780 to 2500 nm and is divided into short-wave near-infrared (wavelength: 780 to 1100 nm) and long-wave near-infrared (wavelength: 1100 to 2500 nm) [6]. It records the overtones and their combinations. of hydrogen-containing groups X-H (C-H, O-H, N-H, etc.) and covers a large amount of structure and composition information, so the components containing these groups can be established by chemometrics with relational models for qualitative or quantitative analysis of samples [4]. The NIRS equipment is relatively simple. The common components include a light source, a beam splitter system, a sample stage, an optical detector, and its data analysis system (Figure 2).

Figure 2. The consists of NIRS measurement.

The NIR spectrometer can be small in size and have portable equipment to facilitate rapid detection in the field, such as testing the quality of agricultural products at the agricultural production site. As for the detection methods, NIRS has transmission, diffuse reflection, transmission, and reflection detection methods, and the choice of different detection methods is also demand-dependent [7].

2.2. Basic Principles and Characteristics of HSI

As for hyperspectral imaging (HSI) technology, it is a detection method that combines spectral technology and image technology, which can obtain both spectral information and image information of samples [6]. The spectroscopic technique mainly involves near-infrared absorption and has been introduced above. Image technology can obtain target image information without touching the object, which has the advantages of intuitive, quantitative, recognition, fast speed, and mature technology.

Compared with NIRS equipment, the structure of hyperspectral equipment is more complex. It mainly includes a CCD camera, imaging spectrometer, lens, light source controller, sample station, mobile platform and its controller, hyperspectral data acquisition software, and mobile platform of mobile control software (Figure 3). The imaging spectrometer is the core component, which is able to obtain both spectral and spatial information of the target. Optical systems need to have higher resolution to ensure the quality of spectra and images. Its detectors usually need to have high sensitivity and high resolution to cope with a large number of spectral bands and fine image information [7] in size and higher in price, and the detectors are often used in aerospace remote sensing and large-scale farmland monitoring in precision agriculture and other fields. The application of hyperspectral imaging technology can not only locate the specific position of the sample, obtain the spectral information of the specific position, and detect the chemical composition of the sample, which is a fast and efficient optical detection means. According to different scanning methods, it can also be divided into point scanning, line scanning, and area scanning, among which line scanning is most commonly used. Table 2 shows the comparison of these technologies.

Figure 3. The consists of HSI measurement (line scan).

Table 2. The comparison between NIRS and HSI.

Technology	Data	Components of the Equipment	Price	Portability or Not
NIRS	Spectral data	Light source, beam splitter system, sample stage, optical detector, and its data analysis system	Lower cost	Yes
HSI	Combination of spectral and spatial data	CCD camera, imaging spectrometer, lens, light source controller, sample station, mobile platform and its controller, data acquisition software, and mobile platform of mobile control software	Higher cost	No

2.3. The Data Processing Methods

As for the processing of NIRS and HSI data, the general analyzing steps are shown in Figure 4. In the processing of NIRS data [6], the general steps (Figure 4A) are divided into sample collection, spectral data preprocessing, feature waveband selection, model establishment and evaluation, etc. At the same time, from the principle of HSI technology, it can be seen that its processing methods include spectral data processing and image data processing (Figure 4B). In general, image data processing and analysis are first carried out, including image preprocessing, image segmentation, feature extraction, etc., and the spectral data processing methods and steps are the same as NIRS data.

Figure 4. The general analyzing steps [6].

When spectral data are collected, the first step is to preprocess the acquired spectra in order to eliminate irrelevant information and noise, such as the electrical noise, sample background noise, and stray light noise, etc. [8]. Common spectral preprocessing methods include smoothing, derivatives, multiple scattering correction (MSC), baseline correction, standard normal transformation (SNV), and combinations of these approaches [9,10,11]. Table 3 shows the description and formula of part of the spectral preprocessing methods.

Table 3. The description and formula of part of the spectral preprocessing methods.

Spectral Preprocessing Method	Description	Formula
Mean Centering	Shifts the spectrum so that the mean of the spectral values is zero.	xnew = x – x0, where x is the original spectral value and x0 is the mean of all spectral values.
SNV	Normalizes the spectra to have a mean of zero and a standard deviation of one.	$x_{SNV}={\displaystyle \frac{x-x_0}{\sigma }}$, where x0 is the mean and $\sigma$ is the standard deviation of the spectral values.
MSC	Corrects for scatter effects in the spectra.	$x_{MSC}={\displaystyle \frac{x-b_0}{b_1}}$, where b0 and b1 are regression coefficients obtained from a linear regression between the sample spectrum and the mean spectrum.
First—Derivative	Calculates the first—order derivative of the spectral data to enhance spectral features.	${\displaystyle \frac{dx}{d\lambda }}$, where $\lambda$ is the wavelength.
Second—Derivative	Calculates the second—order derivative of the spectral data to further enhance spectral features.	${\displaystyle \frac{d^{2}x}{d^2\lambda }}$, where $\lambda$ is the wavelength.

Generally, the second step involves extracting spectral feature wavebands. Since the collected spectral data consist of hundreds or thousands of wavebands, using all wavebands for modeling presents issues such as extensive computation and prolonged processing time. Additionally, due to the lack of distinct spectral absorption, severe overlap, and the inclusion of redundant information, the stability and prediction accuracy of the model may be undermined. Hence, it is common practice to eliminate irrelevant information and filter out independent variables with high correlation during the modeling process. Currently, commonly used feature waveband selection methods [12,13,14] include principal component analysis (PCA), competitive adaptive reweighting (CARS), genetic algorithm (GA), successive projection algorithm (SPA), no-information variable elimination (UVE), etc.

After pretreatment or feature wavelength selection, the calibration model of the spectrum is ultimately established for qualitative or quantitative analysis. With the rapid development of statistics, many inevitably use mathematical analysis methods [15] for more scientific classification and quantitative detection, which can be linear or non-linear, supervised or unsupervised. Common qualitative and quantitative methods include biomimetic pattern recognition (BPR), K-nearest neighbor (KNN), linear discriminant analysis (LDA), partial least squares discriminant analysis (PLS-DA), extreme learning machine (ELM), support vector machine (SVM), backpropagation neural network (BPNN), partial least squares regression (PLSR), radial basis function neural network (RBFNN), etc. In recent years, deep learning algorithms, particularly convolutional neural networks (CNNs), have been utilized for both the quantitative and qualitative modeling of near-infrared spectroscopy [16,17,18]. Compared to traditional machine learning methods, convolutional neural networks can progressively extract features embedded within spectral data through multiple convolutional and pooling layers, thereby reducing the need for extensive pre-processing of spectra and variable selection prior to modeling to some extent [6]. In fact, the efficiency and portability of the model also need to be considered. Some algorithms, such as ELM, BPNN, and RBFNN, may present the disadvantages of slow convergence speed and ease of falling into the local optimal and may be relatively weak in efficiency and portability. However, PLS-DA and PLSR may be more computation-intensive when dealing with large-scale data. The efficiency and portability of BPR algorithm in practical applications may also be affected by the complexity of its implementation. So, this is where the algorithm will improve in the future. After the model is established, it is crucial to evaluate its stability and accuracy and select high-quality models. Commonly used indicators include accuracy, correlation coefficient, the standard deviation of calibration and prediction set samples, etc. Table 4 is the summary of part of the feature wavelength selection and modeling methods and evaluation indicators.

Table 4. The summary of part of the feature wavelength selection and modeling methods and evaluation indicators.

Feature Wavelength Selection Method	Modeling Methods	Evaluation Indicators
PCA, CARS, GA, SPA, UVE, etc.	Traditional methods such as BPR, KNN, LDA, PLS-DA, ELM, SVM, BPNN, PLSR, RBFNN, etc. Deep learning methods such as CNN, etc.	Accuracy, sensitivity, specificity, correlation coefficient, and the standard deviation.

Next, the specific applications of these two technologies in corn variety and purity detection, vitality detection, internal components detection, mycotoxins, and other indicators of detection (freezing damage, hardness detection, and maturity) are introduced. By searching the relevant literature on corn seed quality detection and each indicator as keywords based on NIRS and HSI technology on the Web of Science and CNKI, and pre-sorting the obtained literature, only SCI journal papers and some papers in core Chinese journals were retained for our review. Finally, 33 papers for variety and purity identification, 12 papers for vitality test, 18 for internal component determination, 25 for mycotoxins detection, and 13 for other indicators detection are reviewed.

3. Specific Applications

3.1. Variety and Purity Detection

Regarding variety and purity detection, scholars have conducted relevant research on corn variety identification, transgenic corn detection, and haploid detection (Table A1).

3.1.1. Variety Identification

With the growing global demand for corn and the continuous advancement of breeding technology, the number of new corn seed varieties added each year is on the rise. In North America, parts of Europe and Brazil, Argentina, etc., the number of new varieties is relatively large every year. In around 2010, the number of approved corn varieties in China was in the range of 30 to 40 each year. However, in 2020, the number of approved varieties reached 802 within a single year [19]. Incidents of seed adulteration, counterfeiting, and passing off inferior products as superior ones are becoming increasingly frequent in the market. Unscrupulous merchants mix unqualified seeds with qualified ones, leading to reduced crop yields and seriously undermining the interests of growers. Therefore, variety and purity detection are particularly necessary.

Over the years, many scholars have conducted extensive research in this field. Wu et al. (2010) collected diffuse reflection spectral data of 37 corn varieties. The original spectra were preprocessed by vector normalization, and a method of spectral feature waveband extraction based on standard deviation was proposed. They performed PCA and established a BPR model with an average correct recognition rate of 94.3% [20]. Wang et al. (2011) used genetic algorithms to select feature wavebands for 37 varieties, reducing the data dimension and using linear discriminant analysis for classification, achieving an average correct rejection rate of 99.65% [21]. Huang et al. (2011) used the PLS model to determine the purity of Nongda 108 corn with the final verification set of 95.75% and an average relative error of 2.73% [22]. Jia et al. (2012) established a BPR model for eight corn varieties, with an average correct recognition rate of 94.6% [23]. In 2014, Han et al. collected near-infrared spectral data of eight corn varieties and found that the SVM algorithm was more suitable for small sample spectral analysis under certain principal components [24]. In 2015, Jia et al. used near-infrared spectroscopy and chemometrics to identify coated corn seed varieties, with the soft independent modeling of class analogy (SIMCA) model showing an accuracy rate of 97.5% [25]. In 2018, Cui et al. studied the feasibility of identifying maize seed varieties by combining NIRS with chemometrics methods. The spectra were pretreated with smoothing, the first derivative and vector normalization; then, PCA, LDA, and BPR were applied to establish identification models. The results showed that the average correct identification rate was more than 90%, and it was robust to samples from different regions and years [26].

From 2012 to 2022, various scholars utilized hyperspectral imaging technology. Zhang et al. extracted texture variables (contrast, homogeneity, energy, and correlation) and established a PCA-GLCM-LS-SVM model with a recognition accuracy of 98.89% [27]. Wang et al. collected hyperspectral image data of three varieties (378 samples) of corn seeds over the wavelength region of 400–1000 nm and selected six optimal spectral wavelengths by SPA. Different varieties of corn seeds were classified according to spectral, texture, or fusion data, and the results showed that the classification results based on the DT pretreatment data of the whole band were the best, with an accuracy of 91.667% (Figure 5) [28]. Xia et al. proposed an MLDA algorithm and achieved a classification accuracy of 99.13% [29]. Zhou et al. realized the non-destructive identification of sweet corn seed varieties and achieved good classification accuracy rates [30]. Scholars from Zhejiang University used algorithms like RBFNN, t-SNE, and SVM combined with hyperspectral imaging to classify corn varieties [31,32,33]. Zhao et al., Miao et al., and Bai et al. also made contributions in corn seed variety classification using hyperspectral imaging and different methods [31,32,33].

Figure 5. Data processing: (A) original spectra, (B) preprocessed spectra, (C) optimal wavelengths selected by SPA, and (D) classification images by LS-SVM model based on data fusion [28].

Some scholars proposed model update concepts. He et al. used clustering algorithms for updating the model; the results showed that, after the model parameters were determined and applied, the overall accuracy rate of the updated model was 98.3%, which was higher than the accuracy rate of 84.6% obtained by the non-updated model [34]. With the update and iteration of learning algorithms, deep learning was applied in data analysis [35,36,37,38]. Zhang et al. combined hyperspectral imaging with deep convolutional neural networks (DCNN) to classify four maize seed varieties. The results showed that the DCNN model had 100% training accuracy, 94.4% test accuracy, and 93.3% verification accuracy, which was superior to KNN and SVM models in most cases. The DCNN model also had the best performance in terms of evaluation indicators (sensitivity, specificity, and accuracy) and achieved good results [35]. Wang et al. compared different classification models and found that the performance of the proposed CNN-LSTM is slightly better than the other five models [39].

3.1.2. Transgenic Detection

Transgenic corn is one of the most widely used transgenic crops in China. Although this technology can breed excellent varieties with high yield, high resistance, and high quality that adapt to adverse ecological environments, it also brings two issues: exogenous gene safety and environmental safety. As transgenic technology develops and is widely applied, the safety and reliability of transgenic products continue to draw increasing attention. Accurate, rapid, and efficient detection methods for transgenic corn are of great significance for food quality monitoring and control and for human health and the social economy.

Feng et al. (2018) used NIRS for the identification of transgenic corn seeds. Three variable selection algorithms (weighted regression coefficients, PCA loadings, and second derivatives) were used to extract the feature wavelength. Five methods, including KNN, SIMCA, naive Bayes classifier, ELM, and BPANN, were used to establish the discriminant model. The results showed that ELM exhibited the best performance with 100% classification accuracy based on full waveband and 90.83% based on feature wavelengths [40]. At the same year, Peng et al. utilized NIRS to identify transgenic corn with SG smoothing. The SVM model based on the full spectrum was superior to the PLS model, with a recognition accuracy of over 90% [41]. In 2022, Zhang et al. employed NIRS to identify four groups of transgenic corn. A three-layer ANN model identified transgenic corn with 100% accuracy [42].

In 2017, Feng et al. used NIR hyperspectral imaging and multivariate data analysis along with pretreatment algorithms. Models were established to classify transgenic corn grains with nearly 100% calculation and prediction accuracy [43]. In 2023, Wei et al. collected hyperspectral images of three types of corn grains and compared traditional and deep learning algorithms. The prediction accuracy of the BPNN-GA model was 0.861, and deep learning modeling had an accuracy of 0.961 [44].

3.1.3. Haploid Detection

In the field of identifying corn haploids and polyploids, several studies were conducted. Liu et al. employed KPCA for feature extraction and SVM to establish a classification model for corn seed haploids and polyploids, with average correct recognition rates of 95% and 93.57%, respectively [45]. Yu et al. proposed a non-linear feature analysis method based on SVSKLPP, showing high average accuracy, sensitivity, and specificity of 97.1%, 98.8%, and 95.4% [46]. Cui et al. proposed a screening scheme for corn haploid seeds based on NIRS quantitative analysis with an average accuracy above 90% and a model monitoring and calibration solution for stability [47]. In 2021, Ge et al. fused NMR and NIRS data, improving unclear corn grain category classification by about 9% with a DADA framework [48]. In 2023, Ribeiro et al. used NIRS and preprocessed data with PCA, a PLS-DA model achieved 100% accurate classification of haploid and diploid seeds and plants [49].

As for the application of hyperspectral in haploid detection, Wang et al. took Zhengdan 958 and Nongda 616 as research objects and explored the effect of embryo orientation (embryo facing or facing away from the light source) on the haploid recognition model. The correct acceptance rate of the haploid and diploid test sets was as high as 99%, and the wrong acceptance rate was less than 1% [50]. He et al. discussed the applicability of near-infrared hyperspectral imaging, used three variable selection methods to determine 20 wavelengths and established a PLSDA model with 90.31% accuracy [51]. Zhang et al. used hyperspectral imaging combined with a GAN-based data enhancement method to identify haploid corn grains, showing that both DCGAN and CGAN increased classifier accuracy by more than 10%, with CGAN having a higher increase [52].

3.2. Vitality Detection

Seed vitality is a key indicator for measuring seed quality and is related to germination rate, quantity, and stress resistance. Traditional seed vitality determination methods are destructive and have a long test period. The recent studies are summarized as shown in Table A2.

Agelet et al. used NIRS and various methods (PLS-DA, KNN, and LS-SVM) to identify thermally damaged corn kernels. Among them, PLS-DA had the highest accuracy of 99% [53]. In 2013, Yang et al. used NIRS and BPNN to establish a corn seed vitality detection model. The BPNN model constructed by combining preprocessing and feature extraction methods had a recognition accuracy of 95.0% [54]. In 2018, Li et al. used NIRS to establish a new method for detecting vitality indicators (germination rate, germination index, vitality index, and other vitality indicators) of sweet corn seeds and established a PLSR quantitative model [55]. Wang et al. (2020) distinguished normal seeds from heat-damaged and artificially aged seeds using NIRS based on a self-made single seed device. The accuracy for heat-damaged seeds reached 100%, and it was higher than 95% for artificially aged seeds. The accuracy of the comprehensive model for the calibration set and prediction set were 97.8% and 97.3%, respectively [56]. In 2022, Zhao et al. applied NIRS and chemometrics to determine the vitality of sweet corn seeds under reflection and transmission modes and found that the transmission model was better than the reflection model [57].

As for the applications of HSI technology [58,59,60,61,62,63,64], Ambrose et al. established a PLS-DA method to distinguish aged and normal corn seeds. The classification accuracy rates of the calibration set and prediction set were 97.6% and 95.6%, respectively [58]. In 2018, Wakholi et al. used short-wave infrared hyperspectral imaging (1000–2500 nm) combined with chemometrics methods (LDA, PLS-DA, and SVM) to evaluate corn seed vitality. The results showed that the SVM model combined with the pretreatment methods had an accuracy of up to 100% (Figure 6) [60].

Figure 6. Schematic flow from data collection to final chemical image [60].

In 2022, Cui et al. used regression methods to establish the prediction relationship between hyperspectral features and seedling root length. The coefficient of determination reached 0.8319 [63]. In 2022, Zhao et al. combined hyperspectral imaging and deep convolutional neural network to predict the vitality of waxy corn seeds with an accuracy of 98.83% [62]. These studies show that NIRS and hyperspectral imaging technology have great potential in detecting the vitality of corn seeds. Different methods and models have their own advantages in terms of accuracy.

3.3. Components Determination

The main components of corn include water, protein, starch, and fat. The content level directly affects product prices and market positioning. The internal chemical components of seeds contain chemical bonds such as C-H, H-H, and C-N that are sensitive to near-infrared spectroscopy. Therefore, many scholars have established models for the component detection of seeds. These studies are summarized in Table A3. As can be seen from the table, the spectral range adopted by most studies is in the long wave near-infrared range, because this waveband range can better reflect the chemical composition and internal information of the sample.

3.3.1. Moisture Determination

Moisture is an important factor affecting seed storage, transportation, and germination. Through moisture detection, understanding the moisture content of seeds provides a basis for seed storage and processing. There are some studies conducted in the detection of moisture based on NIRS [65,66,67] and HSI [68,69,70,71,72,73,74]. Fassio et al. (2009) used NIRS to predict high-moisture grain corn’s nutritional value, accurately predicting dry matter, acid detergent fiber, and in vitro organic matter digestibility [65]. Wang et al. (2019) established a PLS monitoring model for corn moisture during filling period with small sample sizes. For 20 and 50 samples, the coefficient of determination was higher than 0.99 [66]. Zhang et al. (2020) collected hyperspectral images of embryo side and endosperm side of seeds and extracted feature wavelengths using UVE. The results showed that the average spectrum extracted from the centroid region was better than that extracted from the whole seed region, and the results of convolutional smoothing pretreatment are better than other pretreatment methods [70]. Wang et al. (2020–2023) explores the accuracy of detecting water content of single maize seed under different pretreatment and waveband extraction algorithms. The results showed that the combination of long-wave hyperspectral imaging technology and the general algorithm model could realize the non-destructive and stable prediction of maize seed water content [69,71,74].

3.3.2. Determination of Other Components

Table A3 shows that most of the studies apply NIRS for composition prediction determination [75,76,77,78,79,80]. Fassio et al. (2015) used NIRS to determine corn seed oil content with coefficient of determination of 0.90%, cross-validation standard error of 0.17%, and RPD of 2.3 for qualitative determination [75]. Lyu et al. (2016) analyzed corn’s crude protein, moisture, and fat [76]. Emmanuel et al. (2022) used NIRS for breeding selection and built an NIRS model for the detection of amino acids [77]. Xu et al. (2023) constructed a BiPLS-PCA-ELM model with the prediction determination coefficients (R²_p) of 0.996, 0.989, 0.974, and 0.976; the prediction root means square errors of 0.018, 0.016, 0.067, and 0.109; the RPD value of 15.704, 9.741, 6.330, and 6.236, respectively, for the prediction of moisture, oil, protein, and starch (Figure 7) [80].

Figure 7. (A) Spectroscopic data of the samples (the average values of the original spectra and the pretreated and normalized moisture, grease, protein, and starch spectra of the sample), (B) characteristic spectral intervals selected by BiPLS, and (C) distribution map of measured and predicted values, (a–d) predicted results of moisture, oil, protein, and starch, respectively [80].

Cataltas et al. (2023) used a one-dimensional convolutional autoencoder and NIRS to detect protein, starch, oil, and moisture [78]. Wu et al. (2023) proposed the use of an A-CARS model, and the results showed that the model was significantly better than other methods [79]. Some studies were conducted in the component determination with HSI technology, Liu et al. (2020) used HSI technology to determine starch content in single corn seeds, smoothing and derivative algorithms were used to preprocess the spectra, and the CARS method was used to select the feature wavelength. The results showed that ANN prediction model based on the Levenberg–Marquardt algorithm was the best model for starch content determination. The correlation coefficient of prediction set was 0.96, and the root mean square error (RMSEP) was 0.98 [81]. Zhang et al. (2022) combined hyperspectral imaging technology and deep learning algorithms (DCGAN, ACNNR, and CNNR) to predict the oil content of a single corn kernel. The results showed that the attention mechanism helped reduce the prediction error, making ACNNR perform best (prediction determination coefficient = 0.9198) [82,83].

3.4. Mycotoxins Detection

Corn seeds may be infected with mycotoxins when not stored properly. If the mycotoxins cannot be detected well in the early stage, it will bring about various hazards. On the one hand, mycotoxins such as aflatoxin and zearalenone seriously endanger human health and animal growth. Transmission through the food chain can lead to problems like liver damage and disorders of the reproductive system. On the other hand, toxins can reduce the quality and nutritional value of corn and affect its economic value. Strict limit standards for mycotoxins require testing to avoid trade disputes and economic losses. At the same time, detection helps farmers take preventive measures and guarantees the sustainable development of agriculture. Table A4 summarizes the research on mycotoxin detection of corn seeds based on NIRS and HSI.

As can be seen from Table A4, whether using NIRS [84,85,86,87,88,89,90] or HSI technology [91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109] for mycotoxin detection, the vast majority of studies are focused on AFB1 detection. Fernández et al. (2009) used NIRS technology to detect the content of aflatoxin B1 in 152 samples. The best predictive model for detecting AFB1 in maize was obtained by using SNVD as scatter correction (r² = 0.80 and 0.82; SECV = 0.211 and 0.200 for grating and FT-NIRS instruments, respectively) [84]. Tallada et al. detected AFB1 using near-infrared reflection spectroscopy and color imaging. The accuracy rates for detecting uninfected and infected kernels were 89% and 79%, respectively [85]. Tao et al. (2019) used NIRS to detect AFB1 contamination on the surface of corn kernels. The results showed that the best three-category model had a prediction overall accuracy of 98.6% in both ranges I and II. For the seven-category discrimination model, the best overall prediction accuracies obtained in ranges I and II were 91.4% and 97.1%, respectively [86]. In the later stage, some scholars independently developed portable near-infrared spectrometers to analyze different levels of AFB1 or fumonisin B1 and B2 [88,89]. Wang et al. (2022) combined NIRS with deep learning algorithms for the determination of AFB1 in corn. The results showed that, compared with the 1D-CNN model, the performance of the 2D-MTF-CNN model was significantly improved [90].

As for the study of HSI technology for aflatoxin detection of corn seeds. Zhu et al. (2016) combined fluorescence and V/NIR HSI to detect aflatoxins in whole corn kernels. The results showed that the best overall prediction accuracy (95.33%) of the LS-SVM model had a threshold of 100 ppb on the embryo side [92]. Conceição et al. (2021) used near-infrared hyperspectral images combined with PLS-DA to rapidly identify Fusarium verticillioides and Fusarium graminearum [93]. Wang’s team from China Agricultural University successively detected AFB1 in single corn kernels from 2014 to 2023 [97,98,99,100,101,102,103,104]. During this period, improvements and applications were made to different corn varieties, wavelength ranges, pretreatments, especially feature extraction and modeling algorithms. Each study showed that hyperspectral imaging was an effective tool for detecting AFB1 in a single corn kernel. Lu et al. (2022) combined short-wave infrared hyperspectral imaging and synchronous Fourier-transform infrared microspectroscopy to study the chemical and spatiotemporal changes in damaged corn kernels caused by Aspergillus flavus infection from macroscopic and microscopic perspectives. For three types of samples, satisfactory full-spectrum models and multispectral models were obtained, respectively, using the PLSR model. In addition, the combination of SR-FTIR microspectroscopy and two-dimensional correlation spectroscopy reveals the possible sequence of dynamic changes in nutrient loss and AFB1 in corn kernels (Figure 8) [106]. Wang et al. (2023) used a fluorescence hyperspectral imaging system for AFB1 detection and developed an undersampling stacking (USS) algorithm for unbalanced data. The results showed that the USS method combined with characteristic wavelength variance analysis achieves the best performance, with an accuracy of 0.98 at a threshold of 20 or 50 μg/kg using the endosperm side spectra [108].

Figure 8. (A) Linear regression plots and pixel–level visualization maps of multispectral models predicted AFB1 levels with the gradual damage on intact (a,d), pierced (b,e) and halved (c,f) samples inoculated by A. flavus over time course. (B) the synchronous (a,c) and asynchronous (b,d) 2DCOS maps generated from the SR-FTIR spectra of inoculated halved samples. (C) SR-FTIR images for chemical distributions of nutrient depletion (a) and trace AFB1 accumulation (b) at significant wavenumbers from the thin section of inoculated halved kernels over inoculated time of 24–96 h [106].

3.5. Other Application Areas

In addition to the above seed quality testing, in the production and application process of corn seeds, it is also necessary to test other qualities such as freezing damage, hardness, and maturity, according to different application scenarios. Freezing damage, hardness, and maturity detection based on NIRS and HSI are summarized in Table A5.

3.5.1. Frost Damage Detection

In cold regions or seasons with changeable climates, corn seeds are prone to suffer different degrees of frost damage, which will reduce the germination rate, growth potential, and yield potential of seeds. Therefore, conducting frost damage detection of corn seeds can help seed production bases detect and deal with problems in time to avoid or reduce losses.

Agelet et al. (2012) tried to use NIRS technology to identify frost-damaged corn seeds. The results showed that this technology could not effectively distinguish frost-damaged corn seeds, and the highest recognition result was only 63.4% [53]. Jia et al. (2016) also tried to identify frost-damaged corn seeds with an initial moisture content of 30% in a low-temperature environment of −19.2 °C. The results showed that this technology could realize the identification of frost-damaged seeds, and the highest average accuracy rate could reach 97% [110]. Zhang et al. (2022) collected near-infrared spectral data of corn seeds under different frost damage conditions and used different preprocessing, feature extraction methods, and modeling methods. The results showed that in the case of standard normal transformation preprocessing combined with the principal component analysis feature extraction method and the K-nearest neighbor model, the classification results of the training set and the test set were 99.4% and 100%, respectively [111].

Zhang et al. (2019) successively used the VIS/NIR hyperspectral imaging system to classify frostbitten corn seeds of different degrees. Three different preprocessing methods (MSC, SNV, and 5–3 smoothing), three wavelength selection algorithms (SPA, PCA, and X-loading), and three modeling methods (PLS-DA, KNN, and SVM) were compared. The results showed that using 5-3 smoothing and the SPA wavelength selection method for modeling could improve the signal-to-noise ratio of the model, and the classification accuracy could reach more than 90% [112]. In 2021, the scholar studied the feasibility of combining hyperspectral imaging (400–1000 nm) with DCNN to classify different frost-damaged corn seeds. For five and four categories of situations, relevant models (KNN, SVM, ELM, and DCNN) were established, and the evaluation indicators (accuracy, sensitivity, specificity, and precision) were compared. The results showed that the accuracy rate of the DCNN model was the most satisfactory (Figure 9) [113].

Figure 9. (A) The average spectra of corn seeds at different freezing conditions, (B) the DCNN structure, where (a) is the DCNN structure, (b) is the Convs block, (c) is the FC block, and (C) visualization map of the seeds with DCNN model [112].

3.5.2. Hardness Detection

Williams et al. (2009, 2016) used near-infrared hyperspectral imaging technology to classify corn seeds of three hardness categories: hard, medium, and soft. The results showed that the sensitivity and specificity based on pixel classification were 0.75 and 0.97, respectively; the model based on score histogram performed better in the classification of hard-grained samples, with a sensitivity and specificity of 0.93 and 0.97, respectively; and the average spectral model had a sensitivity and specificity of 0.95 and 0.93 for medium-sized grains [114,115]. Qiao et al. (2022) collected hyperspectral image data of corn seeds, extracted feature wavelengths by continuous projection algorithm, and established a prediction model of moisture content by PLSR method. Finally, the prediction model was combined with the hardness regression model to verify the hardness prediction model. The results showed that the coefficient of determination of hardness prediction was 0.912 [116].

3.5.3. Maturity Detection

Wang et al. (2015) used hyperspectral imaging technology to predict the texture changes in corn seeds at different storage times. The OSC-SPA-PLSR model was used to visualize the influence of different storage times on texture characteristics in corn seeds. The results showed that the performance of the OSC-PLS full spectral range model in predicting the structural characteristics of corn seeds was significantly better than that of the model without pretreatment [117]. Huang et al. (2016) realized the classification of corn seeds of different years based on hyperspectral imaging and model update [118]. Wang et al. (2022) extracted the average spectra of the embryo side, endosperm side, and both sides. The SVM algorithm was used to develop a classification model based on full spectrum, and PCA was used to extract feature wavelengths. The accuracy rate of the prediction set using the full-spectrum classification model was 100% [119]. Yang et al. (2016) used the band ratio image of 640 nm/525 nm to classify and identify corn seed samples with high and low maturity, with an average correct recognition rate of 93.9% [120]. Wang et al. (2021) collected hyperspectral images of the embryo side and endosperm side of corn seeds of different maturities. The embryo-side spectrum (T1), endosperm-side spectrum (T2), and two-sided fused spectrum (T3) were preprocessed, and feature wavelengths were screened by PCA. The results showed that the PLS-DA classification model established by the 12 feature wavelengths extracted from the T2 spectrum based on SG-first-order derivative pretreatment had the best classification effect, with an average classification accuracy of 100%. When the T1 spectrum was input into the model, its average classification accuracy was 98.7%. The results proved the potential of hyperspectral imaging technology in the rapid and accurate classification of corn seed maturity [121].

4. Challenges and Future Perspectives

4.1. Challenges

Compared with traditional detection methods, spectroscopy and its imaging detection technology have the advantages of easy processing, fast speed, and non-destructiveness. Although NIRS and HSI technologies show great application potential in many fields, they also face a series of challenges and problems. (1) The detection of environmental factors, such as differences in external temperature, humidity, and types of spectrometer components, will lead to differences in the collected spectral data. This directly affects the accuracy of the calibration model, as the compatibility of spectral data is poor, and it is difficult to achieve mutual transplantation. (2) The selection of samples has a significant impact on the reliability and adaptability of the calibration model. If the sample selection is not representative enough, or the concentration range is not widely covered, and the difference in sample color and state leads to a large difference in the obtained NIR spectrum, the adaptability of the model will be reduced. (3) With the continuous progress of spectral imaging technology, the resolution of the obtained data is improved, but, at the same time, a large amount of redundant information will be introduced, increasing the calculation pressure. How to extract useful feature bands and eigenvalues and eliminate irrelevant information has become an urgent problem to be solved. (4) For the calibration model established for corn detection, its robustness and transmission still need to be further improved. Due to the lack of unified industry norms, different instruments are not effective in predicting different varieties and often need to be modeled separately for specific varieties, which hinders the scale and industrial development of near-infrared spectroscopy.

4.2. Future Perspectives

In view of the development status and various problems of spectroscopy-imaging technology in corn detection, the following prospects are put forward in order to improve the application. (1) In the spectrometer production industry, unified specifications and general standards need to be formulated to eliminate cross-instrument obstacles in terms of hardware and software and provide a data interface for development to enhance the portability of models. (2) It is necessary to establish a spectra database of corn and expand the coverage of modeling samples to improve the prediction range and accuracy of corn. (3) Combined with artificial intelligence technology, through deep machine learning, optimal wavebands are selected, and irrelevant information is automatically filtered out without manual screening, and an optimal model is established to enhance the robustness and continuity of the corn quality detection model. (4) Hyperspectral imaging data are usually redundant, which requires effective algorithms to extract feature wavelengths for dimensionality reduction. The application of multiple spectral technologies in agriculture is predicted to be more commonly used in the future, and the characteristics of various technologies will be used to achieve high-quality detection of corn quality. Table 5 shows the details of the contributions and gaps of the previous surveys and this review paper.

Table 5. The contributions and gaps of the previous surveys and this review paper.

Contributions and Gaps	Details
Contributions in previous surveys	(1) Conducted research on NIRS and HSI in the quality evaluation of corn seeds.
Gaps in previous surveys	(1) Not systematically compare and discuss these two technologies; (2) Not summarize the above contributions; (3) Not clearly propose the challenges and future research directions.
Gaps filled by this review paper	(1) Compared and discussed the composition, data collection and processing methods, and portability of these two technologies; (2) Summarized the quality evaluation of corn seeds, including variety purity, vigor, internal components, mycotoxins, as well as other qualities such as freeze damage, hardness, and maturity, emphasized the importance of corn quality characteristics and the functions of applied algorithms; (3) Proposed the challenges and future research directions of spectroscopy and its imaging technology.

5. Conclusions

From the above studies, it can be seen that NIRS and HSI technology have made great progress in corn seed quality detection (variety and purity, vitality, internal components, mycotoxins, and other indicators (freezing damage, hardness detection, and maturity)). The characteristics and principles of these two technologies are introduced, and their components, data acquisition and processing methods, as well as portability are compared and discussed. Breakthroughs and innovations have been made in detection methods, spectral preprocessing methods and recognition algorithms. The significance of corn quality characteristics and the function of the applied algorithm are emphasized. Finally, the challenges and future research direction of spectral and its imaging technology was proposed, aiming to further enhance the accuracy, reliability, and practicability of the detection technology, provide valuable reference information for researchers, and contribute to global food security and sustainable agricultural development.

In summary, with the rapid development of spectrum and its imaging technology, the detection methods of corn quality are also advancing with time. This is not just for corn, but more and more crops can be accurately detected by these technologies. It will become an important means of agricultural production inspection in the future.