Photonic Technologies in Plant and Agricultural Science

Spectral imaging is widely applied in pest detection and yield estimation in agri- and horticulture. Illumination based spectral imaging (instead of selection on the detector side) could provide a more cost-effective solution, potentially enabling applications for which current spectral imaging technology is too costly. In this contribution we present the development of a low-cost approach for quantitative spectral imaging, with spectral selection using LED-based illumination at various wavelengths. We used a Printed Circuit Board (PCB) approach with a motherboard, which is implemented as a Pi-Hat and controls the LEDs. The motherboard is connected to a donut-shaped PCB with 8 surface mounted LEDs with spectra from 400 - 940 nm to select wavelengths. We use a monochrome Si-camera sensitive to all wavelengths, a light diffuser and a 3D printed holder. Using our setup we register a multispectral image by successively switching LED illumination and grabbing frames. We validated the spectral imaging capability using conventional reflection spectroscopy and color tiles. We are currently investigating various applications of this compact, low-cost solution for spectral imaging.

We investigated the viability of using a human laser hair removal system to remove/reduce sheep wool growth to prevent flystrike as an alternative to crutching and mulesing. A Duetto MT EVO laser from Quanta systems, which combines a 755 nm Alexandrite and a 1064 nm Nd:YAG laser was used to treat excised 1 year old lamb skin. Skin samples analysed with histology showed that white wool fibres remained undamaged beneath the epidermis but black wool sustained damage. These results are unsurprising given the complete absence of the laser’s target melanin in white wool and abundance of it in black.

The purity analysis of oilseed rape (Brassica napus L.) is currently a labor-intensive and manual process, requiring significant human effort for accurate assessment. In this context, the KIRa-Sorter system presents an innovative solution that leverages hyperspectral imaging technology for automating the comprehensive classification of various contaminants present in rapeseed samples. The initial phase of the KIRa-Sorter system involves the efficient capture of hyperspectral and RGB image data from rapeseed samples as input for classification. From up to 200 different types of foreign objects typically found in these samples, a reduced coreset has been defined that the system is able to automatically singulate, classify and physically sort.

Ultraviolet-C (UV-C) has been used in food decontamination as an alternative to traditional strategies that are harmful to the food and environment. This study evaluated the effectiveness of UV-C at 222 nm, at doses of 0; 0.90 ± 0.02; 1.70 ± 0.02; 3.40 ± 0.02; 7.0 ± 0.1; 14.0 ± 0.3; 27.5 ± 0.4 mJ/cm², in the decontamination of tomatoes inoculated with Escherichia coli. After application of light, microbiological evaluation and physical analyses (colorimetric parameters, pH and weight) were performed. The microbiological results reaffirmed the effectiveness of UV-C 222 nm in surface decontamination of food. From the results of the microbiological analyzes, good decontamination values were obtained using the light dose of 3.40 ± 0.04 mJ/cm², with a reduction greater than 4 log. However, the physical analyzes did not identify a significant extension in shelf life, showing the absence of damage to the fruit development. In conclusion, UV-C at 222 nm was effective in inactivating E. coli present on tomatoes, without altering their physical characteristics.

Hydroponic cultivation would greatly benefit from continuous monitoring for successful year round and sustainable production. A fast analysis approach for crop growth and nutrient deficiency inspection using hyperspectral imaging (HSI) is discussed. This approach requires minimal computational time, enabling rapid classification of growth stages and timely detection of nutrient stress.

An innovative approach, based on hyperspectral imaging (HSI) coupled with chemometrics, allowing the monitoring by HSI the macro- and micronutrients in vine leaves sampled from different areas of a vineyard, is presented in this study. The proposed approach is based on the acquisition by HSI in the short-wave infrared range (SWIR: 1000-2500 nm) of dried and milled vine leaves, followed by the implementation of a classification model based on Partial Least Square (PLS). Micro-X-ray fluorescence analyses were carried out on the same samples to evaluate the macro- and micronutrients presence and to correlate their contents with the acquired HSI-based spectra.

The color change of oil palm fresh fruit bunches (FFB) is one of the important parameters for harvesting decisions. However, human visual identification is prone to errors due to uncontrolled ambient lighting and the far distance of fruit bunches on tall palm trees. These errors can lead to inaccurate harvesting and significant revenue loss. This study introduces a laser remote sensor for non-destructive FFB ripeness assessment. Based on the unique spectral reflection curves of the FFB at different ripeness, laser modules at three different wavelengths (visible- NIR region) have been employed for the measurement. The photodetector and laser sources are configured in a coaxial manner to enable long work distances up to 9m. This portable laser remote sensor has undergone successful on-plant testing in the oil palm estates, with measurements validated against oil content by conventional bunch analysis. It is a potential tool for precision harvesting and oil yield prediction.

In this study we design a monitoring protocol to inspect the physiological response of a selection of plant species towards fluctuations in environmental variables and pollution levels. We identified two representative areas of the central district and of the peripheral area of the city of Cagliari including different tree and shrub species (Quercus Ilex L.; Nerium Oleander L.; Prunus Cerasifera Ehrh) that are typical of that Mediterranean urban environment. Portable sensing devices based on fluorescence and Raman techniques were employed to non-invasively monitor the levels of main pigments and components of the leaf, which have been related to the characteristics of plant species and of the urban location under examination. Differences among the plant species considered were observed between the two urban zones identified.

Hyperspectral cameras are a key enabling technology in precision agriculture, biodiversity monitoring, and ecological research. Consequently, these applications are fuelling a growing demand for devices that are suited to widespread deployment in such environments. Current hyperspectral cameras, however, require significant investment in post-processing, and rarely allow for live-capture assessments. Here, we introduce a novel hyperspectral camera that combines live spectral data and high-resolution imagery. This camera is suitable for integration with robotics and automated monitoring systems. We explore the utility of this camera for applications including chlorophyll detection and live display of spectral indices relating to plant health. We discuss the performance of this novel technology and associated hyperspectral analysis methods to support an ecological study of grassland habitats at Wytham Woods, UK.

Plant growth is one of the most relevant issues even in our highly technically driven world. Food production comes along with environmental impact through water consumption, use of fertilizer and chemical plant protection. Robotics and machine vision are used to minimize impact. Optical systems in agriculture face tough requirements. Innovative approaches for multi- and hyperspectral imaging systems were developed. Fully reflective optics enable the use of a single objective for a broad spectral range. Together with adequate beam splitters, this allows the use of multiple imaging sensors in only one optical system. The efficient integration of complex non-axial optical systems can be implemented by using a new technology called “place and bend assembly” invented by Fraunhofer IPMS. Now, MEMS based systems will cut down effort furthermore by replacing the NIR hyperspectral camera through a MEMS deflection system, single spot NIR measurement and artificial intelligence (AI) control. Image information from a camera device is used to control the measurement of NIR spectra of selected spots to gather all relevant information required.

The growing demand for innovative agricultural technologies drives our research on photonics-based solutions for non-destructive plant analysis. Our work introduces two techniques to overcome the limitations of conventional methods. The first part presents a portable VIS-NIR handheld spectrometer, showcasing strong correlations (0.84 and 0.77) for quantifying anthocyanin and chlorophyll content. In the second part, a customized fiber optic probe using ATR-FTIR enables measurements of plant parameters (water content, antioxidant activity, lignin, and cellulose). Statistical analysis of high-resolution spectra introduces the novel index NDMRI for effective phytochemical differentiation. These non-destructive, rapid, and objective methods promise to optimize agriculture and drive advancements in this field.

Optical Coherence Tomography (OCT), a non-contact, non-destructive imaging technique, is becoming a popular tool in phytophotonics, helping to address research questions in plant biology and horticulture. However, the stationary nature of typical OCT systems compromises its non-destructive advantage since plants often need to be dissected for an analysis with a laboratory OCT system. Here we present a portable, low-cost OCT system that enables in-situ measurements of plants. We outline technical challenges encountered during the development and showcase initial measurements of different plant tissues.

In this work, we evaluate the capabilities of a portable smartphone-based laser speckle imager for grading seeds based on the extent of biological activity related to germination inside seeds. The portable illumination device is attached to the rear side of a smartphone and externally controlled via USB port. Both RGB and laser speckle image sequences are simultaneously acquired through smartphone cameras for further processing. A custom image processing algorithm is used to generate the final spatio-temporal activity map. Our approach enables better grading of seeds prior to cultivation thereby potentially improving crop yields.

Reduction of pesticide use is one of the most pressing current issues in plant production. Laser weeding could play a significant role in the fundamental changes expected in agriculture. We present this laser process and discuss technical challenges in different agricultural application scenarios. A proper choice of laser technology depends on many factors such as the agricultural machinery it is mounted on as well as on the crop system it is designed for or the type and size of farm that is to use the device on their farmland.

The rising demand for herbicide alternatives in weed management prompts innovative approaches. Laser-based weed control provides precision weeding for individual plants near crops. To be competitive with conventional methods, laser weeding must be optimized for efficiency and effectivity. To optimize and to determine the most successful parameters, plant experiments need to be conducted. Obvious laser parameters to investigate are wavelength, beam diameter, angle of irradiation and treatment point. NIR-wavelengths of 1470 nm and 1940 nm promise high potential for damaging plants due to their absorption by water. A thermal imaging camera aids in visualizing the laser effects, advancing understanding for better weed control practices.

Tracking the freshness of food products throughout the whole supply chain is an increasing interest. Especially the quality of prepackaged products is hard to evaluate for the consumer. We developed a sensor foil, that allows the assessment of packaged foods quality by using color or fluorescence indication. Spoiled food causes the sensor foil to change its color from green to red as well as a change of its fluorescence spectrum. This color change is detectable by using an inexpensive handheld fluorescence spectrometer without opening the package unit. This would enable direct tracking of food freshness along supply chains.

Raman investigations are carried out in selected agri-photonic fields using portable shifted excitation Raman difference spectroscopy (SERDS) sensor systems with implemented 785 nm dual-wavelength diode lasers. Measurements are performed in apple orchards, for on-site soil investigations and the analysis of bovine milk. In the presence of laser induced fluorescence and daylight, SERDS separates Raman signals of target components with an improved signal-to-background-noise ratio by an order of magnitude. This allows a subsequent qualitative and quantitative analysis of selected substances. The results demonstrate the potential of these SERDS sensor systems for, e.g., precision agriculture, soil nutrient management, and on-farm food monitoring.

Detection of Clostridium contamination in milk is of fundamental importance as it leads to blowing defects in cheese. Currently, microbiological detection techniques used in the dairy industry are time-consuming and limited in efficiency and sensitivity. The activity of these bacteria causes the formation of CO2 and H2, due to the fermentation of butyric acid. For this reason, the possible application of a Raman spectroscopy based instrument on headspace gases was investigated. The purpose of this paper is to show the developed technique and the experimental results obtained. Tests on real samples of Clostridium-inoculated milk in vials with culture media revealed a high correlation between the Raman measurement and the reference one (using the most probable number procedure). Clostridium was identified within 24-36 hours of contamination, effectively demonstrating the applicability of the instrument in its context of use. Future developments will focus on instrument engineering for industrial applicability.

Surfaces of terrestrial plant’s aerial organs are enveloped by a polymeric cuticle composed out of cutin, waxes, and polysaccharides. The cuticle protects plant organs against adverse environmental conditions and pathogens. During development, it endures stretching and strain. In apple fruit, continuous synthesis and deposition of cutin and waxes during growth results in an age and strain gradient within the cuticle. To analyze the resulting complex micromechanical structure, a Brillouin scattering setup is used, revealing a strong impact of dewaxing on the mechanical properties and a gradient in the Brillouin frequency shift along the cross section of the fruit skin.

Plant disease outbreaks pose serious threats to global food security. A rigid methodology that accounts for rapid identification of the earliest point of infection caused by plant viruses is necessary. Raman spectroscopy that generates spectral signatures of cellular-level dynamics resonates the virus induced alteration in plants through moderations in spectral features. Here, we present a model study to identify the earliest point of infection. Measured spectra from healthy and virus infected Arabidopsis thaliana plants are applied to principal component analysis. We found a separation as early as 8 days post inoculation between healthy and virus infected plants.

Cyanobacteria, also known as blue-green algae, can produce cyanotoxins which can be harmful to animals and humans and can affect the ecosystem as well as the water quality. In marine or fresh water the cyanobacteria can grow to dense blooms with a high concentration of cells within a few days. Consequently, a fast and ideally real-time observation and analysis of cyanobacterial blooms is very important to ensure safety. We present a Raman spectroscopic approach to investigate and differentiate toxic and non-toxic cyanobacteria. For this, features of the acquired Raman spectra are highlighted to identify harmful cyanobacteria.

Plant in vitro culture techniques are fundamental for research, propagation, and breeding. Automated phenotyping of in vitro cultures can revolutionize trait evaluation by transitioning to continuous and objective quantification, as well as by enhancing accuracy, speed, and efficiency. Limited research exists on automated sensor usage in plant tissue culture, mainly focusing on "plant-to-sensor" approaches. This project developed a novel robotic phenotyping system using low-cost sensors to digitally quantify plant in vitro cultures. Various sensors, including RGB camera, laser distance sensor, micro spectrometer, and thermal camera, were deployed. The system quantified growth and detected physiological disorders using a convolutional neural network.

We present optical methods for contactless measurement of two plant parameters that are relevant for e.g. tomato growth: ‘head thickness’ (the thickness of the stem about 30 cm below the top), and leaf area index (LAI, the leaf area per unit ground area). For LAI a 3D camera is exploited. First results show that leaf area can be determined with an uncertainty <8% under laboratory conditions with would be sufficiently accurate. For head thickness an optical caliper based on a laser and a line array has been developed, showing an uncertainty well below the required 0.1 mm.

Greenhouse production systems are increasingly complex, necessitating a data-driven approach with robust, intelligent sensors. Although there is a clear advantage for growers to be able to monitor the physiological status of the crop, current practices mostly involve cumbersome, expensive, and slow laboratory measurements. It is shown that visible and near-infrared imaging spectroscopy allows for the rapid and non-destructive assessment of the concentrations of sugars, starch, pigments, various nutrients, and dry matter in the leaves and fruits of tomato plants. A tailored feature selection algorithm also shows the feasibility of using as few as 8 bands with spectral cameras. This study validates imaging spectroscopy as a rapid tool for assessing crop status and fruit quality in greenhouse horticulture.

This project aims to develop a hyperspectral remote sensing approach to detect potato virus Y (pathogenic virus of the family Potyviridae, PVY), from an unpiloted aerial vehicle (UAV). The hyperspectral camera is mounted on the UAV to capture the reflectance of the pixels of the leaves and identify the subtle changes in the color as an indicator of the PVY. The PVY-infected plants tend to have visible mosaic patterns on the leaves, leading to a potential signal for optical detection. Managing the PVY is one of the priorities for the Montana Seed Potato Growers, necessitating the development of a rapid-detection system for PVY. We aim to evaluate if we can detect PVY from a UAV with a radiometrically calibrated hyperspectral sensor to measure upwelling radiance and a calibrated spectrometer to measure downwelling irradiance. We aimed to start with publicly available data from Wageningen University, Netherlands, to build a baseline for our model under controlled lighting. However, we encountered difficulty working with this data, and hope to revisit this portion of the effort in the future.

Non-destructive evaluation of polyphenols, present in leaf epidermises and fruit skins, can be determined indirectly by multispectral measurements of the chlorophyll fluorescence from the underlying cell layers. This technique, called the chlorophyll fluorescence excitation screening (ChlFES) method, is based on the spectral attenuation by superficial compounds of the incoming radiation impinging on chlorophyll molecules. Examples of the non-destructive monitoring of plant flavonoids response to different UV radiation regimes are reported, with the use of portable fluorescence sensors. Application of the ChlFES to apples, olives, kiwifruits, plums and mainly wine grape to detect flavonoids and anthocyanins was aimed to select fruits with the highest nutraceutical value.