Robust Classification of Tea Based on Multi-Channel LED-Induced Fluorescence and a Convolutional Neural Network

Toward commercial applications of LED and laser-induced fluorescence techniques for food identity, quality, and safety monitoring: A review

The assessment of food safety and quality is a matter of paramount importance, especially considering the challenges posed by climate change. Convenient, eco-friendly, and non-destructive techniques have attracted extensive attention in the food industry because they can retain food safety and quality. Fluorescence radiation, the process by which fluorophore emits light upon the absorption of ultraviolet or visible light, offers the advantages of high sensitivity and selectivity. The use of excitation-emission matrix (EEM) has been extensively explored in the food industry, but on-site detection of EEMs remain a challenge. To address this limitation, laser-induced fluorescence (LIF) and light emitting diode-induced fluorescence (LED-IF) have been implemented in many cases to facilitate the transition of fluorescence measurements from the laboratory to commercial applications. This review provides an overview of the application of commercially available LIF/LED-IF devices for non-destructive food measurement and recent studies that focus on the development of LIF/LED-IF devices for commercial applications. These studies were categorized into two stages: the preliminary exploration stage, which emphasizes the selection of an appropriate excitation wavelength based on the combination of EEM and chemometrics, and the pre-application stage, where experiments were conducted on scouting with specific excitation wavelength. Although commercially available devices have emerged in many research fields, only a limited number have been reported for use in the food industry. Future studies should focus on enhancing the diversity of test samples and parameters that can be measured by a single device, exploring the application of LIF techniques for detecting low-concentration substances in food, investigating more quantitative approaches, and developing embedded computing devices.

Development of a multispectral fluorescence LiDAR for point cloud segmentation of plants

The accelerating development of high-throughput plant phenotyping demands a LiDAR system to achieve spectral point cloud, which will significantly improve the accuracy and efficiency of segmentation based on its intrinsic fusion of spectral and spatial data. Meanwhile, a relatively longer detection range is required for platforms e.g., unmanned aerial vehicles (UAV) and poles. Towards the aims above, what we believe to be, a novel multispectral fluorescence LiDAR, featuring compact volume, light weight, and low cost, has been proposed and designed. A 405 nm laser diode was employed to excite the fluorescence of plants, and the point cloud attached with both the elastic and inelastic signal intensities that was obtained through the R-, G-, B-channels of a color image sensor. A new position retrieval method has been developed to evaluate far field echo signals, from which the spectral point cloud can be obtained. Experiments were designed to validate the spectral/spatial accuracy and the segmentation performance. It has been found out that the values obtained through the R-, G-, B-channels are consistent with the emission spectrum measured by a spectrometer, achieving a maximum R² of 0.97. The theoretical spatial resolution can reach up to 47 mm and 0.7 mm in the x- and y-direction at a distance of around 30 m, respectively. The values of recall, precision, and F score for the segmentation of the fluorescence point cloud were all beyond 0.97. Besides, a field test has been carried out on plants at a distance of about 26 m, which further demonstrated that the multispectral fluorescence data can significantly facilitate the segmentation process in a complex scene. These promising results prove that the proposed multispectral fluorescence LiDAR has great potential in applications of digital forestry inventory and intelligent agriculture.

Classification of Tea Leaves Based on Fluorescence Imaging and Convolutional Neural Networks

The development of the smartphone and computer vision technique provides customers with a convenient approach to identify tea species, as well as qualities. However, the prediction model may not behave robustly due to changes in illumination conditions. Fluorescence imaging can induce the fluorescence signal from typical components, and thus may improve the prediction accuracy. In this paper, a tea classification method based on fluorescence imaging and convolutional neural networks (CNN) is proposed. Ultra-violet (UV) LEDs with a central wavelength of 370 nm were utilized to induce the fluorescence of tea samples so that the fluorescence images could be captured. Five kinds of tea were included and pre-processed. Two CNN-based classification models, e.g., the VGG16 and ResNet-34, were utilized for model training. Images captured under the conventional fluorescent lamp were also tested for comparison. The results show that the accuracy of the classification model based on fluorescence images is better than those based on the white-light illumination images, and the performance of the VGG16 model is better than the ResNet-34 model in our case. The classification accuracy of fluorescence images reached 97.5%, which proves that the LED-induced fluorescence imaging technique is promising to use in our daily life.

Transient 2D Junction Temperature Distribution Measurement by Short Pulse Driving and Gated Integration with Ordinary CCD Camera

The time resolution of the transient process is usually limited by the minimum exposure time of the high-speed camera. In this work, we proposed a method that can achieve nanosecond temporal resolution with an ordinary CCD camera by driving the LED under test with a periodic short-pulse signal and multiple-cycle superposition to obtain two-dimensional transient junction temperature distribution of the heating process. The temporal resolution is determined by the pulse width of the drive source. In the cooling process, the Boxcar gated integration principle is adopted to complete the two-dimensional transient junction temperature distribution with temporal resolution subject to the minimum exposure time of the CCD camera, i.e., 1 μs in this case. To demonstrate the validity of this method, we measured the two-dimensional transient junction temperature distribution of the blue LEDs according to the principle of thermoreflectance and compared it with the thermal imaging method.

Quantitative Detection of Extra Virgin Olive Oil Adulteration, as Opposed to Peanut and Soybean Oil, Employing LED-Induced Fluorescence Spectroscopy

As it is high in value, extra virgin olive oil (EVOO) is frequently blended with inferior vegetable oils. This study presents an optical method for determining the adulteration level of EVOO with soybean oil as well as peanut oil using LED-induced fluorescence spectroscopy. Eight LEDs with central wavelengths from ultra-violet (UV) to blue are tested to induce the fluorescence spectra of EVOO, peanut oil, and soybean oil, and the UV LED of 372 nm is selected for further detection. Samples are prepared by mixing olive oil with different volume fractions of peanut or soybean oil, and their fluorescence spectra are collected. Different pre-processing and regression methods are utilized to build the prediction model, and good linearity is obtained between the predicted and actual adulteration concentration. This result, accompanied by the non-destruction and no pre-treatment characteristics, proves that it is feasible to use LED-induced fluorescence spectroscopy as a way to investigate the EVOO adulteration level, and paves the way for building a hand-hold device that can be applied to real market conditions in the future.

Confocal hyperspectral microscopic imager for the detection and classification of individual microalgae

We propose a confocal hyperspectral microscopic imager (CHMI) that can measure both transmission and fluorescent spectra of individual microalgae, as well as obtain classical transmission images and corresponding fluorescent hyperspectral images with a high signal-to-noise ratio. Thus, the system can realize precise identification, classification, and location of microalgae in a free or symbiosis state. The CHMI works in a staring state, with two imaging modes, a confocal fluorescence hyperspectral imaging (CFHI) mode and a transmission hyperspectral imaging (THI) mode. The imaging modes share the main light path, and thus obtained fluorescence and transmission hyperspectral images have point-to-point correspondence. In the CFHI mode, a confocal technology to eliminate image blurring caused by interference of axial points is included. The CHMI has excellent performance with spectral and spatial resolutions of 3 nm and 2 µm, respectively (using a 10× microscope objective magnification). To demonstrate the capacity and versatility of the CHMI, we report on demonstration experiments on four species of microalgae in free form as well as three species of jellyfish with symbiotic microalgae. In the microalgae species classification experiments, transmission and fluorescence spectra collected by the CHMI were preprocessed using principal component analysis (PCA), and a support vector machine (SVM) model or deep learning was then used for classification. The accuracy of the SVM model and deep learning method to distinguish one species of individual microalgae from another was found to be 96.25% and 98.34%, respectively. Also, the ability of the CHMI to analyze the concentration, species, and distribution differences of symbiotic microalgae in symbionts is furthermore demonstrated.

Research on the influence of attention and emotion of tea drinkers based on artificial neural network

Tea can help to regulate the mood of human. Based on the influence of tea on people's mood and attention, this study explored the tea concentration when the mood and attention of drinkers are in the best state, and established the best concentration model of tea. Using sampling experiment method to collect objective data, which are then combined with questionnaire survey method to collect subjective data, using the results to establish a neural network algorithm model to test the accuracy of the neural network algorithm model. Experiments show that the correlation coefficient of the output value of the BP neural network model constructed in this study is basically consistent with the actual prediction result. After obtaining data such as age, gender, frequency of tea drinking, and tea drinking concentration of tea drinkers, the constructed back propagation (BP) neural network model can accurately predict the mental state score of tea drinkers. The research will provide certain data support and theoretical basis for the follow-up development of the tea industry. Follow-up work needs to be performed in order to further adjust the scope and accuracy of the control model. Then, a more complete and accurate advanced BP neural network model can be established for different types of tea and other parameters.

An Online Tea Fixation State Monitoring Algorithm Based on Image Energy Attention Mechanism and Supervised Clustering (IEAMSC)

This study aimed at the shortcomings of existing fixation algorithms that are image-based only, and an effective tea fixation state monitoring algorithm was proposed. An adaptive filtering algorithm was used to automatically filter the ineffective information. Using the energy extractor, the complete energy information of each fixation image was extracted. The image energy attention mechanism was used to identify the prominent features, and based on these, the energy data was mapped to generate the data points as the training data. The cluster idea was adopted, and the training data feed the features trainer. The trend center data of the tea processing energy clustering was generated from different color channels. The corresponding decision function was designed which is based on the distance of the cluster center. The fixation degree of each monitoring image set was measured by the decision function. The Euclidean distance of the energy clustering center of the three channels with the same fixation time progressively approached. The triangle formed by these three points had a trend of gradually shrinking, which was first discovered by us. The detection results showed high accuracy compared with the common classification algorithms. It indicates that the algorithm proposed has positive guiding and reference significance.

Processing Fluorescence Spectra for Pollutants Detection Systems in Inland Waters

Development of contaminant detection systems in various natural and industrial environments has been favored in recent years thanks to the evolution of processors and sensors. Our group works specifically on contaminant detection systems in inland waters: immediate and continuous detection is a fundamental requirement in this type of sensing. Regarding the sensors, the proposed system is based on fluorescence, since it offers a method in which there is no contact with water, which means less wear on the components and a great saving in cleaning and maintenance. On the other hand, the spectrum processing is of great importance, since it is used both in the generation of a library of fluorescence spectra taken in the laboratory and in the continuous analysis of the samples and in the comparison algorithm for identification. The validity of the system is based on the last process that is carried out in a very short time. This article describes a system to process spectra in a more accelerated way.

Fluorescence Scheimpflug LiDAR developed for the three-dimension profiling of plants

This work proposes a novel fluorescence Scheimpflug LiDAR (SLiDAR) technique based on the Scheimpflug principle for three-dimension (3D) plant profile measurements. A 405 nm laser diode was employed as the excitation light source to generate a light sheet. Both the elastic and inelastic/fluorescence signals from a target object (e.g., plants) can be simultaneously measured by the fluorescence SLiDAR system employing a color image sensor with blue, green and red detection channels. The 3D profile can be obtained from the elastic signal recorded by blue pixels through elevation scanning measurements, while the fluorescence intensity of the target object is mainly acquired by red and green pixels. The normalized fluorescence intensity of the red channel, related to the chlorophyll distribution of the plant, can be utilized for the classification of leaves, branches and trunks. The promising results demonstrated in this work have shown a great potential of employing the fluorescence SLiDAR technique for 3D fluorescence profiling of plants in agriculture and forestry applications.