Search

The past 50 yr of advances in weed recognition technologies have poised site-specific weed control (SSWC) on the cusp of requisite performance for large-scale production systems. The technology offers improved management of diverse weed morphology over highly variable background environments. SSWC enables the use of nonselective weed control options, such as lasers and electrical weeding, as feasible in-crop selective alternatives to herbicides by targeting individual weeds. This review looks at the progress made over this half-century of research and its implications for future weed recognition and control efforts; summarizing advances in computer vision techniques and the most recent deep convolutional neural network (CNN) approaches to weed recognition. The first use of CNNs for plant identification in 2015 began an era of rapid improvement in algorithm performance on larger and more diverse datasets. These performance gains and subsequent research have shown that the variability of large-scale cropping systems is best managed by deep learning for in-crop weed recognition. The benefits of deep learning and improved accessibility to open-source software and hardware tools has been evident in the adoption of these tools by weed researchers and the increased popularity of CNN-based weed recognition research. The field of machine learning holds substantial promise for weed control, especially the implementation of truly integrated weed management strategies. Whereas previous approaches sought to reduce environmental variability or manage it with advanced algorithms, research in deep learning architectures suggests that large-scale, multi-modal approaches are the future for weed recognition.

A Gabor wavelets based technique was investigated as a potential tool for scene classification (into one of bare patch, plant, or weed) for its ultimate utility in site-specific application of agrochemicals in wild blueberry fields. Images were gathered from five sites located in central Nova Scotia, Canada. Gabor wavelet features extracted from these images were used to classify scenes according to visually determined classes using step-wise linear discriminant analysis. For individual fields, classification accuracy attained ranged between 87.9% and 98.3%; selected Gabor features ranged between 27 and 72; contextual accuracy for herbicide ranged between 67.5% and 96.7%, and contextual accuracy for fertilizer ranged between 63.6% and 97.1%. The pooled scenes yielded a classification accuracy of 81.4%, and contextual accuracy figures of 61.1% and 73.1% for herbicide and fertilizer, respectively, with selected Gabor features of 36. Calibrations based on LDA coefficients from the pooled scenes could help avoid the need to re-calibrate for each field, whereas those based on individual field LDA coefficients could improve accuracy, hence enable saving on expensive agrochemicals.

Since the advent of Huanglongbing [HLB (Candidatus Liberibacter asiaticus)] in Florida, several preliminary reports have emerged about the positive effects of mineral nutrition on the performance of HLB-affected citrus (Citrus sp.) trees. HLB-affected trees are known to undergo significant feeder root loss. Therefore, studies have focused on foliar nutrient application instead of soil-applied nutrients speculating that the HLB-affected trees root systems may not be competent in nutrient uptake. Some studies also suggest that HLB-affected trees benefit from micronutrients at higher than the recommended rates; however, the results are often inconclusive and inconsistent. To address this, the goal of the present study was to evaluate the nutrient uptake efficiency and the quantitative and qualitative differences in nutrient uptake of HLB-affected trees. HLB-affected and healthy sweet orange (Citrus sinensis) trees were grown in a 100% hydroponic system with Hoagland solution for 8 weeks. The trees were deprived of any fertilization for 6 months before the transfer of trees to the hydroponic solution. Altogether, the four treatments studied in the hydroponic system were healthy trees fertilized (HLY-F) and not fertilized (HLY-NF), and HLB-affected trees fertilized (HLB-F) and not fertilized (HLB-NF). HLY-F and HLY-NF trees were found to have similar levels of leaf nutrients except for N, which was found to be low in nonfertilized trees (HLY and HLB). Both HLB-F and HLB-NF trees had lower levels of Ca, Mg, and S compared with HLY trees. In addition, HLB-NF trees had significantly lower levels of micronutrients Mn, Zn, and Fe, compared with HLY-NF trees. The hydroponic solution analysis showed that HLB-F and HLY-F trees had similar uptake of all the nutrients. Considering that HLB-affected trees have a lower root-to-shoot ratio than healthy trees, nutrient uptake efficiency per kilogram of root tissue was significantly higher in HLB trees compared with HLY trees. Under nutrient-deficient conditions (day 0) only nine genes were differentially expressed in HLB roots compared with HLY roots. On the other hand, when fertilizer was supplied for ≈1 week, ≈2300 genes were differentially expressed in HLB-F roots compared with HLY-F roots. A large number of differentially expressed genes in HLB-F were related to ion transport, root growth and development, anatomic changes, cell death, and apoptosis compared with HLY-F trees. Overall, anatomic and transcriptomic analyses revealed that HLB-affected roots undergo remarkable changes on transitioning from no nutrients to a nutrient solution, possibly facilitating a high uptake of nutrients. Our results suggest the roots of HLB-affected trees are highly efficient in nutrient uptake; however, a small root mass is a major limitation in nutrient uptake. Certain micronutrients and secondary macronutrients are also metabolized (possibly involved in tree defense or oxidative stress response) at a higher rate in HLB-affected trees than healthy trees. Therefore, a constant supply of fertilizer at a slightly higher rate than what is recommended for micronutrients and secondary macronutrients would be beneficial for managing HLB-affected trees.

Leaching of irrigation water from blueberry (Vaccinium spp.) plants intensifies when sandy soils are amended with pine (Pinus spp.) bark. In a greenhouse study, leaching fractions of water (LFW) and nutrients (LFN) were determined from two blueberry (V. corymbosum L.) cultivars, 'Emerald' and 'Jewel', grown in biochar-treated and nontreated sandy subsoil and irrigated with drip emitters using one of three pulse frequencies in a factorial design. The LFW was 50% under Emerald and 20% under Jewel, which has a more extensive root system. When the frequency of same volume of irrigation was increased from 2 (F2) to 10 (F10) pulses per day, the average LFW decreased from 46 to 30%. The LFN from a single fertigation was much lower than the total LFN after 6 d of irrigations applied between two fertigation events. The LFN increased linearly with LFW. Amending subsoil with 2% biochar increased soil pH, limited root growth, and did not reduce nutrient leaching. Eighteen months after plant establishment, 1.9% of applied NO

Huanglongbing is a citrus disease that reduces yield, crop quality, and eventually causes tree mortality. The putative causal agent, Candidatus Liberibacter asiaticus (Rhizobiales: Rhizobiaceae), is vectored by the Asian citrus psyllid, Diaphorina citri Kuwayama. Disease management is largely through vector control, but the insect is developing pesticide resistance. A nonchemical approach to vector management is to grow citrus under screen cages either as bags over individual trees or enclosures spanning many acres. The enclosing screen reduces wind, alters temperature relative to ambient, and excludes a variety of pests that are too large to pass through the screen. Here we evaluated the potential of six screens to exclude D. citri. We conclude that screens with rectangular openings need to limit the short side to no more than 384.3 µm with a SD of 36.9 µm (40 mesh) to prevent psyllids from passing through the screen. The long side can be at least 833 µm, but the efficacy of screens exceeding this value should be tested before using in the field.

Agricultural herbicide application efficiency can be improved using smart sprayers which provide site-specific, rather than broadcast, applications of agrochemicals. The YOLOv3-Tiny convolutional neural network (CNN) was trained to detect two weeds, hair fescue and sheep sorrel, in images captured from wild blueberry fields throughout Nova Scotia, Canada. An evaluation was performed in three commercial wild blueberry fields in Nova Scotia to examine the effects of camera selection and target distance on detection accuracy. A Canon T6 DSLR camera, an LG G6 smartphone, and a Logitech c920 webcam were used to capture RGB images at varying distances from target weeds. Mean F1-scores for each combination of camera and image height were analysed in a 3 × 3 factorial arrangement for hair fescue and a 3 × 2 factorial arrangement for sheep sorrel. Images captured from 0.98 m with the LG G6 and Canon T6 produced F1-scores of up to 0.97 for detection of at least one hair fescue tuft. Images captured with the LG G6 and Canon T6 DSLR from 0.57 m achieved F1-scores of 0.94 and 0.93, respectively, for detection of at least one sheep sorrel plant per image. Sheep sorrel was undetectable in images from the Logitech c920 under 19 of 27 parameter combinations. Future work will involve using the CNN to control herbicide applications with a real-time smart sprayer. Additionally, the CNN will be used in a web-based application to detect target weeds and provide site-specific information to aid management decisions. Using a CNN to detect weeds will create improvements in management techniques, resulting in cost-savings and greater sustainability for the wild blueberry industry.

The wild blueberry crop requires the heavy application of agrochemicals for the proper crop yield and weed control. An integrated machine vision based weed detection system was developed to target goldenrod weed spot-specifically. Color co-occurrence matrices and statistical classifiers were used for the goldenrod detection. The linear and quadratic classifiers were developed using different reduced sub-sets of textural features. The classifiers were evaluated using accuracy, specificity, sensitivity, and false negative ratio at Laboratory scale. Performance of the developed weed detection system was also evaluated in two wild blueberry fields. The results indicated that quadratic classifier with DM-HSISD model showed the best performance at laboratory scale with the classification accuracies of 94.98% and 93.80% for training and testing datasets, respectively. Another quadratic classifier (with DM-HSI model) containing all 39 textural features was found to be the second best with the accuracies of 94.29% and 91.46% on training and test datasets, respectively. The linear classifiers didn’t perform well compare to their respective quadratic counter-parts. Field performance of the developed system equipped with the quadratic DM-HSISD classifier indicated no significant difference between variable rate (VR) and uniform application (UA) in terms of mean percentage area coverage (PAC) for the targeted goldenrod spots in both fields. However, a significant difference was observed between mean PAC of VR and UA applications for the non-targeted wild blueberry spots. The potential and actual chemical savings were in ranges between 46.71% and 74.83% and 30.12% and 60.58% depending on the weed and sprayed area, respectively. These results demonstrated that the developed weed detection system has potential for targeted application of agrochemicals to control goldenrod in wild blueberry fields.

Weed control between plastic covered, raised beds in Florida vegetable crops relies predominantly on herbicides. Broadcast applications of post-emergence herbicides are unnecessary due to the general patchy distribution of weed populations. Development of precision herbicide sprayers to apply herbicides where weeds occur would result in input reductions. The objective of the study was to test a state-of-the-art object detection convolutional neural network, You Only Look Once 3 (YOLOV3), to detect vegetation both indiscriminately (1-class network) and to detect and discriminate three classes of vegetation commonly found within Florida vegetable plasticulture row-middles (3-class network). Vegetation was discriminated into three categories: broadleaves, sedges and grasses. The 3-class network (Fscore = 0.95) outperformed the 1-class network (Fscore = 0.93) in overall vegetation detection. The increase in target variability when combining classes increased and potentially negated benefits from pooling classes into a single target (and increasing the available data per class). The 3-class network Fscores for grasses, sedges and broadleaves were 0.96, 0.96 and 0.93 respectively. Recall was the limiting factor for all classes. With consideration to how much of the plant was identified (broadleaves and grasses), the 3-class network (Fscore = 0.93) outperformed the 1-class network (Fscore = 0.79). The 1-class network struggled to detect grassy weed species (recall = 0.59). Use of YOLOV3 as an object detector for discrimination of vegetation classes is a feasible option for incorporation into precision applicators.

Objective We aimed to determine the rate of Candidatus Liberibacter asiaticus (CLas) DNA disappearance from Huanglongbing (citrus greening disease)-affected citrus trees after bacterial death from heat treatment. Results Using the leaf disc sampling method, we followed CLas qPCR Ct values in leaves of potted ‘Valencia’ orange trees after a heat-treatment that eliminated viable CLas. Although titer declined, CLas remained detectable 5 months after treatment. These results warn caution in interpreting experiments that use CLas titer to assess effects therapeutic treatments on diseased plants. Residual DNA may be detectable for 6 months, but the time to decay depends on pretreatment titer.

BACKGROUND Weed infestations reduce turfgrass aesthetics and uniformity. Postemergence (POST) herbicides are applied uniformly on turfgrass, hence areas without weeds are also sprayed. Deep learning, particularly the architecture of convolutional neural network (CNN), is a state-of-art approach to recognition of images and objects. In this paper, we report deep learning CNN (DL-CNN) models that are remarkably accurate at detection of broadleaf weeds in turfgrasses. RESULTS VGGNet was the best model for detection of various broadleaf weeds growing in dormant bermudagrass [Cynodon dactylon (L.)] and DetectNet was the best model for detection of cutleaf evening-primrose (Oenothera laciniata Hill) in bahiagrass (Paspalum notatum Flugge) when the learning rate policy was exponential decay. These models achieved high F1 scores (>0.99) and overall accuracy (>0.99), with recall values of 1.00 in the testing datasets. CONCLUSION The results of the present research demonstrate the potential for detection of broadleaf weed using DL-CNN models for detection of broadleaf weeds in turfgrass systems. Further research is required to evaluate weed control in field conditions using these models for in situ video input in conjunction with a smart sprayer. © 2019 Society of Chemical Industry.

Precision spraying of herbicides can significantly reduce herbicide use. The detection system is the critical component within smart sprayers that is used to detect target weeds and make spraying decisions. In this work, we report several deep convolutional neural network (DCNN) models that are exceptionally accurate at detecting weeds in bermudagrass [Cynodon dactylon (L.) Pers.]. VGGNet achieved high F1 score values (>0.95) and out-performed GoogLeNet for detection of dollar weed (Hydrocotyle spp.), old world diamond-flower (Hedyotis cormybosa L. Lam.), and Florida pusley (Richardia scabra L.) in actively growing bermudagrass. A single VGGNet model reliably detected these summer annual broadleaf weeds in bermudagrass across different mowing heights and surface conditions. DetectNet was the most successful DCNN architecture for detection of annual bluegrass (Poa annua L.) or Poa annua growing with various broadleaf weeds in dormant bermudagrass. DetectNet exhibited an excellent performance for detection of weeds while growing in dormant bermudagrass, with F1 scores >0.99. Based on the high level of performance, we conclude that DCNN-based weed recognition can be an effective decision system in the machine vision subsystem of a precision herbicide applicator for weed control in bermudagrass turfgrasses.

Building a physical barrier around citrus trees or groves can prevent contact between trees and Asian citrus psyllids, the carrier of the pathogen that causes huanglongbing. This new 4-page article is for growers, scientists, and industries that are interested in this approach to protect citrus from huanglongbing. Written by Timothy A. Ebert, Arnold W. Schumann, Laura Waldo, Daniel Stanton, and Philippe Rolshausen, and published by the UF/IFAS Department of Soil and Water Sciences.

Optimal harvest windows for wild blueberries (Vaccinium angustifolium Ait.) can be as small as a few days to ensure peak ripeness. Confirming this critical timing is one of the most important steps in harvesting wild blueberries. If the harvest window is mistimed it can result in a significant portion of the berries being under or overripe and ultimately unmarketable. In this study, four different convolutional neural networks (YOLOv3, YOLOv3-Tiny, YOLOv3-SPP and YOLOv4) were trained on the Darknet deep learning framework and compared. Each of the networks were trained to recognize green (unripe), red (unripe) and blue (ripe) berries from a series of 6,766 labelled images. These images were cropped from 337 high resolution images taken across four commercial wild blueberry field sites on eight days during July and August of 2018 and 2019. Independent testing yielded the best results with the YOLOv3-SPP network. YOLOv3-SPP achieved AP scores of 88.40% (blue), 78.91% (green), 66.37% (red) for a mAP score of 77.90%. Data from this study in combination with yield data will be integrated into a mobile phone app which will be available for determining harvest timings and estimating yields.

This study evaluated the potential of using machine vision in combination with deep learning (DL) to identify the early blight disease in real-time for potato production systems. Four fields were selected to collect images (n = 5199) of healthy and diseased potato plants under variable lights and shadow effects. A database was constructed using DL to identify the disease infestation at different stages throughout the growing season. Three convolutional neural networks (CNNs), namely GoogleNet, VGGNet, and EfficientNet, were trained using the PyTorch framework. The disease images were classified into three classes (2-class, 4-class, and 6-class) for accurate disease identification at different growth stages. Results of 2-class CNNs for disease identification revealed the significantly better performance of EfficientNet and VGGNet when compared with the GoogleNet (FScore range: 0.84–0.98). Results of 4-Class CNNs indicated better performance of EfficientNet when compared with other CNNs (FScore range: 0.79–0.94). Results of 6-class CNNs showed similar results as 4-class, with EfficientNet performing the best. GoogleNet, VGGNet, and EfficientNet inference time values ranged from 6.8–8.3, 2.1–2.5, 5.95–6.53 frames per second, respectively, on a Dell Latitude 5580 using graphical processing unit (GPU) mode. Overall, the CNNs and DL frameworks used in this study accurately classified the early blight disease at different stages. Site-specific application of fungicides by accurately identifying the early blight infected plants has a strong potential to reduce agrochemicals use, improve the profitability of potato growers, and lower environmental risks (runoff of fungicides to water bodies).

The uniform application (UA) of agrochemicals results in the over-application of harmful chemicals, increases crop input costs, and deteriorates the environment when compared with variable rate application (VA). A smart variable rate sprayer (SVRS) was designed, developed, and tested using deep learning (DL) for VA application of agrochemicals. Real-time testing of the SVRS took place for detecting and spraying and/or skipping lambsquarters weed and early blight infected and healthy potato plants. About 24,000 images were collected from potato fields in Prince Edward Island and New Brunswick under varying sunny, cloudy, and partly cloudy conditions and processed/trained using YOLOv3 and tiny-YOLOv3 models. Due to faster performance, the tiny-YOLOv3 was chosen to deploy in SVRS. A laboratory experiment was designed under factorial arrangements, where the two spraying techniques (UA and VA) and the three weather conditions (cloudy, partly cloudy, and sunny) were the two independent variables with spray volume consumption as a response variable. The experimental treatments had six repetitions in a 2 × 3 factorial design. Results of the two-way ANOVA showed a significant effect of spraying application techniques on volume consumption of spraying liquid (p-value < 0.05). There was no significant effect of weather conditions and interactions between the two independent variables on volume consumption during weeds and simulated diseased plant detection experiments (p-value > 0.05). The SVRS was able to save 42 and 43% spraying liquid during weeds and simulated diseased plant detection experiments, respectively. Water sensitive papers’ analysis showed the applicability of SVRS for VA with >40% savings of spraying liquid by SVRS when compared with UA. Field applications of this technique would reduce the crop input costs and the environmental risks in conditions (weed and disease) like experimental testing.

Visual identification of nutrient deficiencies in foliage is an important diagnostic tool for fine-tuning nutrient management of citrus. This new 2-page article describes a new smartphone app that uses a trained neural network to identify disease and pest symptoms on citrus leaves through your phone's camera. Written by Arnold Schumann, Laura Waldo, Perseveranca Mungofa, and Chris Oswalt, and published by the UF/IFAS Department of Soil and Water Sciences.https://edis.ifas.ufl.edu/ss691

This new 2-page article provides instructions for using the Diagnosis and Recommendation Integrated System, or DRIS, a web tool designed for analyzing leaf nutrient concentrations of Florida citrus. Written by Arnold Schumann and published by the UF/IFAS Department of Soil and Water Sciences.

Goosegrass is a problematic weed species in Florida vegetable plasticulture production. To reduce costs associated with goosegrass control, a post-emergence precision applicator is under development for use atop the planting beds. To facilitate in situ goosegrass detection and spraying, tiny- You Only Look Once 3 (YOLOv3-tiny) was evaluated as a potential detector. Two annotation techniques were evaluated: (1) annotation of the entire plant (EP) and (2) annotation of partial sections of the leaf blade (LB). For goosegrass detection in strawberry, the F-score was 0.75 and 0.85 for the EP and LB derived networks, respectively. For goosegrass detection in tomato, the F-score was 0.56 and 0.65 for the EP and LB derived networks, respectively. The LB derived networks increased recall at the cost of precision, compared to the EP derived networks. The LB annotation method demonstrated superior results within the context of production and precision spraying, ensuring more targets were sprayed with some over-spraying on false targets. The developed network provides online, real-time, and in situ detection capability for weed management field applications such as precision spraying and autonomous scouts.

The information provided in the 2008 2nd edition is still sound for healthy citrus trees under Florida production conditions. Much of the information provided in this document on nutrients, application methods, leaf and soil sampling and irrigation scheduling are also effective for huanglongbing (HLB) affected citrus trees. However, research conducted since HLB was detected in Florida in 2005 has established changes in many production practices, including nutrient rates, irrigation scheduling, soil pH management, and use of Citrus Under Protective Screen (CUPS). Changes to the 2nd edition of SL253 will appear in boxes similar to this one at the beginnings of chapters 2, 6, 8, 9, and 11.

The information provided in the 2008 2nd edition is still sound for healthy citrus trees under Florida production conditions. Much of the information provided in this document on nutrients, application methods, leaf and soil sampling and irrigation scheduling are also effective for huanglongbing (HLB) affected citrus trees. However, research conducted since HLB was detected in Florida in 2005 has established changes in many production practices, including nutrient rates, irrigation scheduling, soil pH management, and use of Citrus Under Protective Screen (CUPS). Changes to the 2nd edition of SL253 will appear in boxes similar to this one at the beginnings of chapters 2, 6, 8, 9, and 11.

To achieve optimal grove nutrition, citrus growers must test grove soil before beginning any fertilization program. Standard procedures for sampling, preparing, and analyzing soil should be followed for meaningful interpretations of the test results and accurate recommendations. This new two-page fact sheet, published by the UF/IFAS Department of Soil and Water Sciences, provides illustrated soil sampling procedures and tables to aid in basic interpretation of lab results. Written by Davie Kadyampakeni, Kelly Morgan, Arnold Schumann, and Rhuanito S. Ferrarezi.https://edis.ifas.ufl.edu/ss667

Preemergence herbicides are typically applied by broadcasting to the top of raised beds before laying the plastic mulch in plasticulture production systems. Broadleaf and grass emergence is limited to transplant holes in the mulch. As a result, most herbicides are applied under the mulch in locations where weeds cannot emerge and herbicides are unnecessary. To reduce this excessive off-target application, a precision hole-punch sprayer was developed at the University of Florida for use in plasticulture production systems. The technology facilitates the application of herbicides during the hole-punch operation immediately before transplant. Application of napropamide and S-metolachlor in an application volume of 233 L ha–1 of water using the precision hole-punch applicator had no effect on tomato and bell pepper growth and yield. Equipment accuracy ranged from 55% to 90%. Preemergence herbicide use was reduced by 88% to 92% with no reduction in weed control. The hole-punch applicator is an effective way to reduce PRE herbicide use in transplant vegetables grown using the plasticulture production system.Nomenclature: Napropamide; S-metolachlor; bell pepper, Capsicum annuum L.; tomato, Solanum lycopersicum L.

Weed interference during crop establishment is a serious concern for Florida strawberry [Fragaria×ananassa(Weston) Duchesne ex Rozier (pro sp.) [chiloensis×virginiana]] producers. In situ remote detection for precision herbicide application reduces both the risk of crop injury and herbicide inputs. Carolina geranium (Geranium carolinianumL.) is a widespread broadleaf weed within Florida strawberry production with sensitivity to clopyralid, the only available POST broadleaf herbicide.Geranium carolinianumleaf structure is distinct from that of the strawberry plant, which makes it an ideal candidate for pattern recognition in digital images via convolutional neural networks (CNNs). The study objective was to assess the precision of three CNNs in detectingG. carolinianum. Images ofG. carolinianumgrowing in competition with strawberry were gathered at four sites in Hillsborough County, FL. Three CNNs were compared, including object detection–based DetectNet, image classification–based VGGNet, and GoogLeNet. Two DetectNet networks were trained to detect either leaves or canopies ofG. carolinianum. Image classification using GoogLeNet and VGGNet was largely unsuccessful during validation with whole images (FscoreG. carolinianumdetection during validation for VGGNet (Fscore=0.77) and GoogLeNet (Fscore=0.62). TheG. carolinianumleaf–trained DetectNet achieved the highestFscore(0.94) for plant detection during validation. Leaf-based detection led to more consistent detection ofG. carolinianumwithin the strawberry canopy and reduced recall-related errors encountered in canopy-based training. The smaller target of leaf-based DetectNet did increase false positives, but such errors can be overcome with additional training images for network desensitization training. DetectNet was the most viable CNN tested for image-based remote sensing ofG. carolinianumin competition with strawberry. Future research will identify the optimal approach for in situ detection and integrate the detection technology with a precision sprayer.

Models help to describe and predict complex processes and scenarios that are difficult to understand or measure in environmental management systems. Thus, model simulations were performed (i) to calibrate HYDRUS-2D for water and solute movement as a possible decision support system for Candler and Immokalee fine sand using data from microsprinkler and drip irrigation methods, (ii) to validate the performance of HYDRUS-2D using field data of microsprinkler and drip irrigation methods, and (iii) to investigate Br, NO, and water movement using annual or seasonal weather data and variable fertigation scenarios. The model showed reasonably good agreement between measured and simulated values for soil water content ( = 0.87-1.00), Br ( = 0.63-0.96), NO-N ( = 0.66-0.98), P ( = 0.25-0.78), and K ( = 0.44-0.99) movement. The model could be successfully used for scheduling irrigation and predicting nutrient leaching for both microsprinkler and drip irrigation systems on Florida's sandy soils.

Objective This study was carried out with the purpose of testing the ability of deep learning machine vision to identify microscopic objects and geometries found in chemical crystal structures. Results A database of 6994 images taken with a light microscope showing microscopic crystal details of selected chemical compounds along with 180 images of an unknown chemical was created to train and test, respectively the deep learning models. The models used were GoogLeNet (22 layers deep network) and VGG-16 (16 layers deep network), based on the Caffe framework (University of California, Berkeley, CA) of the DIGITS platform (NVIDIA Corporation, Santa Clara, CA). The two models were successfully trained with the images, having validation accuracy values of 97.38% and 99.65% respectively. Finally, both models were able to correctly identify the unknown chemical sample with a high probability score of 93.34% (GoogLeNet) and 99.41% (VGG-16). The positive results found in this study can be further applied to other unknown sample identification tasks using light microscopy coupled with deep learning machine vision.

Wild blueberry crop yields are dependent on heavy agrochemical applications to control weeds competing with the crop. Goldenrod is a creeping herbaceous perennial weed that occurred in 90% of wild blueberry fields surveyed in Nova Scotia. A colour co-occurrence matrices (CCMs) based algorithm can be used for spot-application of herbicide on goldenrod within wild blueberry fields. The objective of this study was to analyse the effect of different parameters on computational complexity and goldenrod identification accuracy of CCMs in wild blueberry cropping system. An image acquisition graphical user interface (GUI) and CCMs based textural analysis algorithm were developed in Microsoft Visual® C# programming language. The GUI was used to acquire 2244 area of interest images along with a set of 2244 full frame images from two different wild blueberry fields in central Nova Scotia, Canada. Six image intensity levels, seven image sizes, and three CCMs were used to for the study. The results indicated that intensity levels and image size significantly influenced the computational requirements of CCMs. Images with 256 or 128 intensity levels could be used as these levels correctly classified 94% and 89% of test observations respectively. The processing times were increased from 535 μs to 10,650 μs and 9864 μs to 63,750 μs as the intensity increased from 8 to 256 levels and image size increased from 16 × 16 to 1024 × 1024 pixels, respectively. The time required for textural feature extraction was not statistically significant for different image sizes used in this study. Overall, the results indicated that intensity levels of 128 or 256, a unit image size of 128 × 128 pixels, and saturation colour plane alone or in combination with hue can help to minimise the processing burden without compromising the classification accuracy for real-time applications.

An essential part of the wild blueberry cropping system is the proper management of agrochemical inputs including herbicides, fungicides and insecticides. A machine vision system was developed and mounted on the rear sprayer boom 0.18 m in front of the sprayer nozzles capable of targeting the agrochemical application on an as-needed basis. The three-point hitch mounted sprayer featured 27 nozzles over a 13.7 m boom width and a storage tank capacity of 1135 l. Nine digital color cameras continually take images in real-time while computer software processes the images in 0.15 s to determine the target locations where the nozzles open and spray at speeds up to 1.77 m s−1. Two wild blueberry fields in central Nova Scotia were selected for smart sprayer performance testing with spot-application (SA) of agrochemical as compared to control and uniform application techniques. Chateau® herbicide was applied in a field with an infestation of hair cap moss. Spray droplet comparison showed moss patches were properly targeted using the smart sprayer. SA provided the same coverage performance as compared to uniform on the moss targets with herbicide application savings of 78.5% using the smart sprayer. Harvestable yield results were similar for all application tracks. TruPhos Magnesium and ZincMax foliar fertilizers were tank mixed with Bravo® and Proline® fungicides and applied to compare the difference of SA, control and uniform application. Results showed SA of foliar fertilizer and fungicide led to less premature leaf drop and increased the blueberry stem height, number of branches, stem diameter and fruit buds. SA of foliar fertilizer and fungicide also increased the percent of healthy wild blueberry plants by 57.8% and the harvestable yield by 137.8%. Fungicide application savings using the smart sprayer for SA were 11.6%.

Northeastern North America is the world’s leading producer of wild blueberry. Ripening of wild blueberry is the leading factor for fruit quality. Currently, there are no protocols available for the farming community related to wild blueberry fruit ripening and maturity. A nondestructive, rapid, and reliable digital photography technique could be used to examine the ripening of wild blueberries for appropriate harvesting time. Two wild blueberry fields were selected to examine the berry ripening levels using digital photographic techniques at different time of harvest (early, middle, and late seasons). The fields were divided into four blocks and each block was further divided into three classes of times of harvest. Fruit images from each block were acquired and processed to count blue pixels from each image. A significant correlation was found between percentage of blue pixels and actual fruit yield in Field A (R2 = 0.96; P < 0.001) and Field B (R2 = 0.97; P < 0.001). The results also indicated that the effect of time of harvest on fruit yield was significant and fruit yield increased gradually during early harvesting, reached maximum during mid-season, and then started to decrease in late harvesting. Results indicated that 90% of green-berries had turned blue at the end of middle season compared to early season (58%). Based on the results of this study, optical analysis could help to keep fruit quality by optimizing appropriate harvesting time of wild blueberries. It is also suggested that the optimum time to harvest wild blueberries is middle season to ensure high fruit yield and quality. Keywords: Blue pixels, Fruit yield, Harvesting season, Wild blueberry.

Completely enclosed screen houses can physically exclude contact between the asian citrus psyllid [ACP (Diaphorina citri)] and young, healthy citrus (Citrus sp.) trees and prevent huanglongbing (HLB) disease development. The current study investigated the use of antipsyllid screen houses on plant growth and physiological parameters of young ‘Ray Ruby’ grapefruit (Citrus ×paradisi) trees. We tested two coverings [enclosed screen house and open-air (control)] and two planting systems (in-ground and container-grown), with four replications arranged in a split-plot experimental design. Trees grown inside screen houses developed larger canopy surface area, canopy surface area water use efficiency (CWUE), leaf area index (LAI) and LAI water use efficiency (LAIWUE) relative to trees grown in open-air plots (P < 0.01). Leaf water transpiration increased and leaf vapor pressure deficit (VPD) decreased in trees grown inside screen houses compared with trees grown in the open-air plots. CWUE was negatively related to leaf VPD (P < 0.01). Monthly leaf nitrogen concentration was consistently greater in container-grown trees in the open-air compared with trees grown in-ground and inside the screen houses. However, trees grown in-ground and inside the screen houses did not experience any severe leaf N deficiencies and were the largest trees, presenting the highest canopy surface area and LAI at the end of the study. The screen houses described here provided a better growing environment for in-ground grapefruit because the protective structures accelerated young tree growth compared with open-air plantings while protecting trees from HLB infection.

Precision herbicide application can substantially reduce herbicide input and weed control cost in turfgrass management systems. Intelligent spot-spraying system predominantly relies on machine vision-based detectors for autonomous weed control. In this work, several deep convolutional neural networks (DCNN) were constructed for detection of dandelion (Taraxacum officinale Web.), ground ivy (Glechoma hederacea L.), and spotted spurge (Euphorbia maculata L.) growing in perennial ryegrass. When the networks were trained using a dataset containing a total of 15,486 negative (images contained perennial ryegrass with no target weeds) and 17,600 positive images (images contained target weeds), VGGNet achieved high F1 scores (≥0.9278), with high recall values (≥0.9952) for detection of E. maculata, G. hederacea, and T. officinale growing in perennial ryegrass. The F1 scores of AlexNet ranged from 0.8437 to 0.9418 and were generally lower than VGGNet at detecting E. maculata, G. hederacea, and T. officinale. GoogleNet is not an effective DCNN at detecting these weed species mainly due to the low precision values. DetectNet is an effective DCNN and achieved high F1 scores (≥0.9843) in the testing datasets for detection of T. officinale growing in perennial ryegrass. Moreover, VGGNet had the highest Matthews correlation coefficient (MCC) values, while GoogleNet had the lowest MCC values. Overall, the approach of training DCNN, particularly VGGNet and DetectNet, presents a clear path toward developing a machine vision-based decision system in smart sprayers for precision weed control in perennial ryegrass.