{"id":19993,"date":"2025-04-19T03:38:44","date_gmt":"2025-04-19T00:08:44","guid":{"rendered":"https:\/\/vastraholding.com\/en\/?p=19993"},"modified":"2025-04-19T03:43:49","modified_gmt":"2025-04-19T00:13:49","slug":"drone-uv-vis-plant-stress-detection","status":"publish","type":"post","link":"https:\/\/vastraholding.com\/en\/drone-uv-vis-plant-stress-detection\/","title":{"rendered":"Early Detection of Plant Stress Using UV\u2013VIS Drone Imaging"},"content":{"rendered":"\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Drone-Based UV\u2013VIS Imaging Technology for Early Detection of Plant Stress<\/h1>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t
\n\t\t\t\n\t\t\t\t\t\t<\/span>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Hartmut K. Lichtenthaler (1996) defined plant stress as any adverse condition or factor that disrupts a plant\u2019s metabolism, growth, or development. This definition indicates that stress can stem from internal factors like hormonal imbalances or external ones such as temperature fluctuations, salinity, or the presence of pathogens, leading to diverse physiological responses.<\/p>\n

Plant stress is generally categorized into two main types: abiotic and biotic. Abiotic stress is caused by non-living factors such as drought, salinity, extreme temperatures, heavy metals, and environmental pollutants. Biotic stress, on the other hand, results from living threats like pests, fungi, bacteria, viruses, and competition from neighboring species. Accurately understanding the specifics of each type is essential for selecting effective management strategies.<\/p>\n

When facing stress, plants exhibit complex physiological and biochemical responses, including changes in stomatal conductance, accumulation of free radicals, synthesis of heat shock proteins (HSPs), and modulation of hormonal signaling pathways. For instance, a drop in water pressure triggers the release of abscisic acid (ABA), leading to stomatal closure and reduced transpiration to minimize water loss.<\/p>\n

The economic impact of plant stress is substantial. According to the FAO, up to 40% of global crop production is lost annually due to pests and plant diseases, resulting in over $220 billion in economic losses. Moreover, studies show that abiotic stress factors like drought and unsuitable temperatures can reduce crop yields by more than 60%, posing a serious threat to food security.<\/p>\n

Early detection of plant stress\u2014before visual symptoms appear\u2014enables the implementation of corrective measures such as optimizing irrigation schedules, applying targeted fertilizers, or managing pathogens effectively. Delayed detection, however, can lead to rapidly escalating damage and costly recovery efforts. Therefore, the development of early warning systems is critical for farmers.<\/p>\n

Traditional methods for stress detection, such as visual field inspections and laboratory tissue analysis, are often time-consuming, labor-intensive, and costly. Due to their episodic nature and delayed data processing, these methods are not suitable for continuous monitoring over large areas or for capturing plant health changes in real time.<\/p>\n

To overcome the limitations of conventional methods, advanced non-destructive technologies like remote sensing\u2014utilizing spectroscopy sensors and multispectral cameras\u2014now enable widespread, continuous monitoring of plant health. These tools detect subtle changes in plant reflectance and calculate indices such as NDVI and PRI to assess plant vitality.<\/p>\n

UV\u2013VIS imaging captures light reflectance across specific wavelengths, allowing researchers to track changes in pigment composition, such as chlorophyll and carotenoids. These changes occur before any visible discoloration appears, making them valuable indicators for early detection of water and nutrient stress.<\/p>\n

Integrating UV\u2013VIS imaging with drone platforms allows for the collection of high-resolution data at timely intervals. This approach not only covers extensive agricultural areas but also significantly improves the speed and accuracy of plant monitoring by reducing the need for constant on-site access.<\/p>\n

In summary, plant stress is one of the major threats to sustainable food production, with effects spanning from cellular to economic and societal levels. Early pre-symptomatic detection using innovative, non-invasive technologies\u2014particularly drone-based UV\u2013VIS imaging\u2014can play a vital role in minimizing environmental risks and optimizing resource management in agriculture.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t

\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\"Early\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Principles of UV\u2013VIS Imaging and Biometric Indices<\/h2>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t
\n\t\t\t\n\t\t\t\t\t\t<\/span>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

– Physical Principles of UV\u2013VIS Light Absorption and Reflection<\/h3>\n

In the UV\u2013VIS spectral range, which spans approximately from 300 to 700 nanometers, plants exhibit different behaviors in terms of light absorption and reflection. Chlorophyll pigments strongly absorb blue and red light while reflecting green light. Meanwhile, near-infrared (NIR) light is strongly reflected by the cellular structure of leaves. This contrast in reflectance allows UV\u2013VIS imaging sensors to detect changes in pigment composition and plant structure by capturing reflected light intensities across specific bands.<\/p>\n

Additionally, the ultraviolet (UV) band\u2014ranging from 300 to 400 nanometers\u2014triggers the accumulation of phytochemicals such as flavonoids and phenolics in response to environmental stressors like drought and oxidative stress. These compounds serve as protective agents against UV radiation, increasing UV reflectance as their concentrations rise. Thus, comparing reflectance values in the UV and VIS bands offers deeper insight into the plant\u2019s defensive mechanisms.<\/p>\n\n

– Key Spectral Indices for Assessing Plant Health<\/h3>\n

The Normalized Difference Vegetation Index (NDVI) is one of the most widely used indicators, utilizing reflectance values from the red (RED) and near-infrared (NIR) bands. NDVI is calculated using the formula (NIR\u2013RED)\/(NIR+RED), yielding values between \u20131 and +1. Values close to +1 indicate dense, healthy vegetation, while values below 0 suggest non-vegetative surfaces.<\/p>\n

The Photochemical Reflectance Index (PRI), introduced by John Gamon and colleagues, monitors changes in the xanthophyll cycle by measuring reflectance at 531 and 570 nanometers. PRI is sensitive to the carotenoid-to-chlorophyll ratio and is considered a valuable indicator of photosynthetic light-use efficiency as well as plant response to light and water stress. PRI values typically range from \u20130.2 to +0.2.<\/p>\n

Beyond NDVI and PRI, indices such as the Red Green Ratio Index and the Normalized Difference Water Index (NDWI) are also used to assess nutritional status and water stress. For example, NDWI compares reflectance from NIR and shortwave infrared (SWIR) bands to estimate leaf and soil moisture content.<\/p>\n\n

– Integrating Biometric Indices for Early Stress Detection<\/h3>\n

Using multiple spectral indices simultaneously can significantly enhance the sensitivity and accuracy of early stress detection. Recent studies have shown that combining NDVI and PRI with machine learning algorithms allows for the precise differentiation between normal and drought-stressed conditions with over 90% accuracy. Furthermore, including UV-based indices in the data input improves the system\u2019s ability to detect oxidative stress.<\/p>\n

In drone-based systems, rapid real-time transmission of spectral data to a central processor enables AI-driven inference algorithms to generate high-resolution stress maps across entire fields. These maps assist farmers in making informed management decisions, such as scheduling irrigation or fertilization, while also reducing operational costs.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Implementing Drone Imaging Systems in Precision Agriculture<\/h2>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t
\n\t\t\t\n\t\t\t\t\t\t<\/span>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

– Sensor Selection and Auxiliary Equipment<\/h3>\n

Choosing the right sensor is one of the first and most critical steps in deploying a drone imaging system. Depending on the monitoring objective\u2014such as early detection of water stress, analyzing vegetation density, or identifying pests\u2014RGB, multispectral, or hyperspectral cameras may be used. Multispectral cameras<\/strong> typically collect data in red, green, blue, and near-infrared (NIR) bands and are ideal for indices like NDVI, while hyperspectral cameras<\/strong> provide access to over a hundred spectral bands, enabling more complex diagnostics such as distinguishing between different stress types.<\/p>\n

In addition to the sensor itself, auxiliary equipment such as a gimbal, high-precision GNSS module, and a data logger play a significant role in image quality and geospatial accuracy. For instance, a 3-axis gimbal reduces unwanted vibrations during flight, enhancing image clarity. An RTK-enabled GNSS module can improve positional accuracy to within 5 centimeters, which is essential for creating detailed farm maps.<\/p>\n\n

\u2013 Pablo J. Zarco-Tejada, Professor of Precision Agriculture and Remote Sensing at the University of Melbourne:<\/cite> “It is essential that drone platforms are capable of carrying hyperspectral and thermal technologies to enable non-visual and early detection of plant stress.”<\/blockquote>\n

Selecting a suitable drone also depends on the sensor’s weight and the mission duration. Fixed-wing drones are efficient for covering large areas, but for shorter flights and low-altitude maneuvering, multi-rotor types (quad- or hexacopters) are more commonly used. Battery capacity and the ability to quickly swap batteries in the field become particularly important when multiple sequential flights are required.<\/p>\n\n

– Flight Planning and Imaging Scenarios<\/h3>\n

Accurate flight planning includes defining the drone\u2019s altitude, speed, and flight pattern to ensure uniform coverage of the target area. Flight altitude is chosen based on the required spatial resolution. For instance, to achieve a Ground Sample Distance (GSD) of 5 cm, a flight height of about 50 meters is necessary. While flying higher increases area coverage per image frame, it also reduces spatial resolution.<\/p>\n

The most common pattern for agricultural imaging missions is the “lawnmower” pattern, which scans the field in parallel lines step by step. Flight planning software like Pix4Dfields or DJI Terra allows for image overlap settings\u2014typically 70% for width and 60% for length\u2014to minimize the need for Ground Control Points (GCPs) during post-processing and reduce the risk of data gaps in the final map.<\/p>\n

To improve early stress detection accuracy, periodic flights can be scheduled at fixed intervals\u2014such as weekly or every ten days. This enables continuous monitoring of vegetation changes and allows for the extraction of time-series trends in biometric indices, triggering alerts before significant damage occurs.<\/p>\n\n

– Data Processing and Generating Plant Health Maps<\/h3>\n

After the drone mission is completed, the collected sensor data undergoes either local or cloud-based processing. In the first stage, specialized software (such as Agisoft Metashape or Pix4Dmapper) is used to automatically calibrate, stitch, and transform RGB or multispectral images into orthomosaic maps. In the next stage, spectral indices such as NDVI, PRI, and NDWI are calculated and mapped onto the imagery.<\/p>\n

For more advanced analysis, machine learning algorithms can be employed. Regression or classification models (e.g., Random Forest or SVM) can use spectral index values and textural image features as inputs to differentiate between stressed and healthy areas with over 90% accuracy. For example, combining NDVI and PRI in a Random Forest model has achieved 93% accuracy in detecting drought stress conditions.<\/p>\n

Finally, the generated maps are delivered as GIS layers to agricultural experts, enabling them to optimize irrigation and nutrient schedules through spatial analysis and timely decision-making. This approach can reduce water and chemical input usage by up to 20%, while increasing crop yield by as much as 15%.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Image Data Processing and Stress Detection Algorithms<\/h2>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t
\n\t\t\t\n\t\t\t\t\t\t<\/span>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

– Image Preprocessing and Radiometric Calibration<\/h3>\n

To ensure accuracy in analyzing drone-based UV\u2013VIS imagery, preprocessing steps are essential and include radiometric and geometric corrections, noise reduction, and calibration. Initially, reflectance reference panels are used to convert raw digital values into actual reflectance data, compensating for ambient light variations and camera settings. Next, using Ground Control Points (GCPs) and orthomosaic algorithms, the images are transformed into geo-referenced maps. This process guarantees spatial and spectral consistency across all pixels, making them directly comparable.<\/p>\n\n

– Extracting Spectral and Spatial Features<\/h3>\n

Once calibration is complete, feature extraction becomes a key step in preparing data for plant stress detection. By calculating vegetation indices<\/strong> like NDVI and PRI, along with UV-based custom indices, anomalous reflectance patterns can be identified. Additionally, texture analysis<\/strong> techniques applied to hyperspectral images allow extraction of spatial features such as leaf heterogeneity and canopy density, which are crucial in detecting subtle stress-induced changes. Combining spectral indices with textural features significantly improves the accuracy of distinguishing between healthy and stressed areas\u2014often exceeding 90%.<\/p>\n\n

\u2013 Pablo J. Zarco-Tejada, Spectral Index Researcher and Stress Detection Algorithm Specialist:<\/cite> \u201cIt is essential that agricultural image processing algorithms not only achieve high accuracy but also offer interpretable, physically meaningful outputs.\u201d<\/blockquote>\n\n

– Implementing Machine Learning and Deep Learning Algorithms<\/h3>\n

For stress detection, both machine learning<\/strong> algorithms (such as Random Forest and SVM) and deep learning<\/strong> methods (like convolutional neural networks) are utilized to analyze multiple features simultaneously. Random Forest, using an ensemble of decision trees, automatically assigns weights to spectral and spatial features for classification. Studies show that integrating NDVI and PRI into a Random Forest model can achieve over 93% accuracy in detecting drought stress. Meanwhile, CNN models fed with raw hyperspectral images can learn complex spatial patterns and further enhance detection of damaged zones.<\/p>\n

After classification, the results must be validated using cross-validation<\/strong> techniques and compared with ground-truth data such as stem water potential or leaf health indices. For example, in a study on walnut orchards, a Random Forest model combining thermal and spectral indices successfully predicted stem water potential with a mean absolute error (MAE) of 0.80 bar.<\/p>\n

To communicate the results to farmers, algorithm outputs are typically integrated into GIS platforms and presented as thematic maps. Many cloud-based systems also support real-time processing and periodic updates, enabling farm managers to track stress trends over time and make well-informed management decisions.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\"Early\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Case Studies, Challenges, and Future Outlook<\/h2>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t
\n\t\t\t\n\t\t\t\t\t\t<\/span>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

– Practical Examples in Vineyards and Orchards<\/h3>\n

In studies conducted on Merlot vineyards in a semi-arid region of Spain, Luz C. Atencia Payares and colleagues utilized drone-based thermal imaging and the plant-air temperature differential index (Tc\u2013Ta) to assess water stress levels. This simple index outperformed more complex methods like CWSI in accuracy and provided reliable insights into vine moisture status at different times of day, enabling more efficient irrigation planning.<\/p>\n

In peach and nectarine orchards, researchers Blenort J., Marsal J., and Girona J. successfully tracked seasonal changes in water stress using drone-based thermal imagery. Their findings showed that high-resolution thermal maps could reveal subtle variations in moisture across the growing season, supporting targeted sub-zone irrigation management within orchards.<\/p>\n

Additionally, in citrus orchards in southern France, Gonz\u00e1lez-Dugo and Zarco-Tejada combined the CWSI index with near-infrared (NIR) reflectance data to demonstrate a strong correlation between water stress levels and crop yield. This non-invasive approach enabled early estimation of water requirements and potential economic outcomes.<\/p>\n

Another study on poplar species at European research stations showed that drone thermal imaging could differentiate genetic drought stress responses within a single species. Researchers found that about 25% of drought-resistant clones exhibited significantly lower canopy temperatures compared to sensitive varieties, detectable even before visual symptoms appeared.<\/p>\n\n

– Technical and Operational Challenges in Drone Deployment<\/h3>\n

One of the key technical challenges is selecting and integrating imaging sensors with the appropriate drone platform. Multispectral and hyperspectral cameras often weigh over 1.5 kilograms, requiring high-capacity fixed-wing or multirotor drones, which can significantly increase startup costs.<\/p>\n

Beyond sensor weight, limited battery life (typically 20 to 40 minutes) and long recharge times reduce the area that can be covered per flight. For example, a standard multirotor drone may need more than two flights to cover just one hectare, and in remote or rugged areas, safe landing sites may be hard to access.<\/p>\n

Weather conditions also affect data quality. Strong winds can cause image distortion, and sunlight intensity at different times of day requires repeated camera recalibration. While radiometric correction algorithms and reference reflectance panels can improve accuracy, they demand time and skilled personnel.<\/p>\n

Large volumes of raw data (often hundreds of gigabytes per mission) can strain local storage and processing infrastructure. Many farmers and consultants lack adequate computing resources and must rely on cloud services, which are limited by internet upload speeds.<\/p>\n

Finally, national aviation regulations in some countries restrict drone operations by altitude, proximity to urban areas, or mandatory permits. These rules can delay mission planning and reduce operational flexibility for users.<\/p>\n\n

– Future Outlook and Emerging Technologies<\/h3>\n

One emerging trend is the integration of drones with Internet of Things (IoT) networks and ground-based sensors to create hybrid monitoring systems. In such systems, soil, weather, and imaging data are continuously aggregated on a cloud platform, where artificial intelligence algorithms process them in real time.<\/p>\n

Edge computing\u2014processing data directly onboard the drone\u2014enables the execution of lightweight machine learning models, reducing network load by transmitting only key results. For instance, next-generation systems may autonomously detect stress zones and adjust flight paths accordingly without user intervention.<\/p>\n

Convolutional neural networks (CNNs) and deep learning models are also enabling image analysis without relying on predefined indices. Architectures like U-Net can process multispectral and thermal imagery to generate highly detailed maps of stressed areas, and even identify specific pests or diseases.<\/p>\n

The advancement of autonomous drone swarms capable of simultaneous flights across large areas promises faster coverage, cost reduction, and higher operational efficiency. Additionally, as hyperspectral CMOS sensors become more compact and affordable, they are expected to be integrated into lightweight drones.<\/p>\n

Standardizing calibration protocols, building open data repositories, and publishing validation methods will allow researchers and farmers to compare results across regions and rapidly update their practices. Ultimately, combining UV\u2013VIS imaging with thermal infrared and LiDAR technologies offers a comprehensive view of plant health and structure, greatly enhancing intelligent water and input management in agriculture.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"

UV\u2013VIS drone-based imaging enables the detection of subtle changes in leaf pigments and structure before any visible symptoms appear, helping to improve irrigation efficiency.<\/p>\n","protected":false},"author":1,"featured_media":20042,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[275,126],"tags":[],"class_list":["post-19993","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-digital-ag-iot","category-vastra-article"],"_links":{"self":[{"href":"https:\/\/vastraholding.com\/en\/wp-json\/wp\/v2\/posts\/19993","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/vastraholding.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/vastraholding.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/vastraholding.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/vastraholding.com\/en\/wp-json\/wp\/v2\/comments?post=19993"}],"version-history":[{"count":0,"href":"https:\/\/vastraholding.com\/en\/wp-json\/wp\/v2\/posts\/19993\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/vastraholding.com\/en\/wp-json\/wp\/v2\/media\/20042"}],"wp:attachment":[{"href":"https:\/\/vastraholding.com\/en\/wp-json\/wp\/v2\/media?parent=19993"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/vastraholding.com\/en\/wp-json\/wp\/v2\/categories?post=19993"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/vastraholding.com\/en\/wp-json\/wp\/v2\/tags?post=19993"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}