제한된 훈련 데이터셋에서 이미지 변환 및 2단계 전이학습을 사용한 향상된 식물 병증 인식

허명락 전북대학교 일반대학원 2023 국내박사

식량 안보는 현재 주요 사회 문제 중 하나로 관련 기술은 지난 수십 년 동안 크게 개선되었다. 그러나 유엔에 따르면, 전 세계적으로 약 10\%의 인구가 기아로 고통받고 있으며, 2021년도에 인구의 3분의 1이 충분한 영영을 섭취하지 못하고 있다. 또한 많은 나라에서 농업과 관련된 노동자의 수가 감소하고 있다. 한국의 경우, 농업에 종사하는 인구는 2014년에는 6\%, 2020년에는 전체 인구의 약 4.5\%만을 차지하는 데 그쳤다. 더구나, 기타 많은 요인이 작물의 성장 및 생산량 증가에 위험이 된다. 본 논문에서 연구자는 작물 질병에 초점을 맞추어 연구를 수행한다. 다른 요소들과 비교했을 때, 작물 질병의 처치가 현재 기술과 사회적 경향을 고려할 때 식량 문제를 완화하는 쉬운 방법의 하나일 것이기 때문이다. 농업 전문가의 지식 및 전통적 기계 학습 기반의 작물 질병 인식 방법과 비교하여 딥러닝을 사용한 방법은 최근 몇 년 동안 유용한 결과를 달성하였다. 그러나 문제는 원하는 성능을 얻기 위해서 많은 양의 데이터가 필요하다는 점이다. 그리고 실제 시나리오와 많은 애플리케이션에서 이미지를 수집하고 주석을 설정하는 것은 복잡하고 비용이 많이 든다. 따라서 제한된 훈련 데이터 세트만으로 딥러닝을 사용하여 더 높은 성능을 얻는 것이 중요한 연구 분야로 인식되었다. 본 논문은 이러한 도전 과제 하에서 작물 질병 인식에 대한 딥 러닝의 수행을 촉진하는 것을 목표로 한다. 이 문제를 해결하기 위해 본 연구자는 작물 질병 인식을 그 특성을 가진 컴퓨터 비전 작업으로 간주하였으며, 이러한 관점에서 style-consistent image translation 알고리즘과 transfer learning 전략을 제안한다. 건강한 잎 이미지가 질병에 걸린 잎보다 높은 intra-class variation에서 수집되기가 훨씬 쉽다는 관찰에서 영감을 받아, 건강한 잎을 질병이 있는 잎으로 translation 하지만, 이미지의 스타일은 건강한 잎에서 질병이 있는 잎으로 높은 intra-class variation을 취할 수 있는 스타일 손실 함수를 사용하여 유지된다. 한편, 우리는 일반적인 컴퓨터 비전의 데이터 세트와 작물 질병의 차이를 발견했다. 일반적이고 합리적인 가정은 비슷한 source 및 target 데이터 세트가 transfer learning 패러다임에 유익하다는 것이다. 따라서 2,885,052개의 이미지와 80,000개의 클래스를 가진 작물 관련 데이터 세트인 PlantCLEF2022이 모델을 pre-train 하는 데 사용되었다. PlantCLEF2022의 train 시간을 줄이기 위해 ImageNet에서 pre-trained된 공개 모델을 PlantCLEF2022에서 한 번 더 pre-triain 한 후 target 데이터 세트에서 fine-tuning 하는 2단계 transfer learning 방법을 제안하였다. 마지막으로 알고리듬을 증명하기 위한 실험을 구현하였다. style-consistent image translation은 5개의 클래스와 999개의 이미지를 포함한 토마토 질병 데이터 세트에서 검증된다. data augmentation 및 translation이 없는 이전 이미지와 비교하여, 우리의 알고리듬은 6개의 이미지 분류, 4개의 객체 감지 및 2개의 인스턴스 분할 모델에서 이미지 분류와 객체 감지 및 인스턴스 분할 모두에서 우수한 결과를 달성한다. 질적으로 본 연구에 사용된 딥러닝 모델은 여러 종류의 토마토 질병 잎을 생성하였고, 그 차이를 이해할 수 있었다. 또한, 2단계 transfer learning 전략은 13개의 작물 질병 관련 데이터 세트에서 검증되었다. 우리의 방법을 사용한 성능 평가를 살펴보면, 8개의 데이터 세트 설정에서 일반적으로 채택되는 transfer 학습을 확실한 차이로 능가하였다. 또한 2단계 transfer learning은 20-short 사례에서 12개 데이터 세트에 대해 현재의 state-of-the-art 성능인 73.53\%보다 12.76\% 높은 86.29\%의 유의미한 평균 테스트 정확도 향상을 달성하였다.

Antagonistic Effects of Bacillomycins, 32,33-didehydroroflamycoin and Lucensomycin against Anthracnose and Gray Mold as Cell Envelope Disruptors Identified from Microbial Secondary Metabolites

Kim, Jeong Do 고려대학교 대학원 2016 국내박사

A microbial metabolite library was screened for antifungals that disrupt cell envelope of plant pathogenic fungi by using adenylate kinase (AK)-based cell lysis assay. Bacillus sp. strain GYL4, Streptomyces sp. strain DY46 and Streptomyces sp. strain CA5 were selected for harboring potent AK-releasing activity in their culture extract. B. amyloliquefaciens strain GYL4 was isolated from leaf tissue of pepper plants (Capsicum annuum L.). The antifungal compounds disrupting fungal cell envelope produced by strain GYL4 were identified and also, endophytic feature of strain GYL4 was examined. An eGFP-expressing strain of GYL4 (GYL4-egfp) was constructed and reintroduced into pepper plants, which confirmed its ability to colonize the internal tissues of pepper plants. GYL4-egfp was observed in the root and stem tissues 4 days after treatment and abundantly found in the internal leaf tissue 9 days after treatment. Anthracnose symptoms were markedly reduced in the leaves of pepper plants colonized by GYL4. Bacillomycin derivatives as the active components of GYL4 displayed control efficacy on anthracnose development in cucumber (Cucumis sativus L. cv. Chunsim). S. rectiviolaceus strain DY46 was isolated from soil samples obtained from Danyang province. The cell extract of the strain DY46 markedly reduced disease incidence (by 20.0%) of gray mold on tomato fruits. 32,33-didehydroroflamycoin (DDHR) was identified from the culture extract as an antifungal component which showed inhibitory effects against the mycelial growth of various plant pathogenic fungi at concentrations of 8 to 64 µg mL-1. DDHR completely inhibited gray mold development on tomato fruits at a concentration of 1000 µg mL-1. DDHR (100 µg mL-1) reduced disease incidence (by 11.1%) of gray mold on tomato fruits inoculated with a concentration of 104 conidia mL-1. Also, DDHR showed short residence time (2 days) in tomato fruits treated with DDHR. S. plumbeus strain CA5 was isolated from soil samples obtained from Chuncheon province. The cell extract of the strain CA5 markedly reduced disease incidence (by 22.2%) of gray mold on grapes. Lucensomycin was identified as the active component which inhibited the mycelial growth of various plant pathogenic fungi at concentrations of 1 µg mL-1. Lucensomycin completely inhibited gray mold development on grapes at a concentration of 100 µg mL-1. These results shows that microbial metabolites with AK releasing activity would be a plenteous reservoir of fungicidal compounds for plant disease control, especially for post harvest disease control that require short-term residual activity along with potent efficacy. Adenylate kinase (AK) 어세이를 사용하여 미생물 대사산물 라이브러리로부터 식물병원진균의 세포외피를 파괴하는 항균활성 물질을 탐색하였으며 배양 추출물에서 강한 AK 방출 활성을 보여주는 Bacillus sp. strain GYL4과 Streptomyces sp. strain DY46 그리고 Streptomyces sp. strain CA5 균주를 선발하였다. B. amyloliquefaciens strain GYL4 균주는 고추식물의 잎 조직으로부터 분리하였으며 GYL4 균주가 생산하는 진균의 세포외피를 파괴하는 항균활성 물질을 동정하고 또한 GYL4 균주의 내생적 특성을 조사하였다. eGFP를 발현하는 GYL4-egfp 균주를 구축하고 고추 식물 내부 조직에 정착하는 능력을 확인하기 위해서 고추 식물에 재도입하였다. GYL4-egfp 균주는 처리 후 4일 후에 뿌리 및 줄기 조직에서 관찰되었고 9일 후에 잎 조직에서 발견되었다. GYL4 균주가 정착한 고추 잎에서 탄저병 병징이 현저하게 감소하였다. GYL4 균주의 유효 성분인 bacillomycin 파생물질들은 오이 탄저병에 대한 방제 효과를 보여주었다. S. rectiviolaceus strain DY46 균주는 단양 지역에서 수집한 토양으로부터 분리하였다. DY46 균주의 세포 추출물은 토마토 잿빛 곰팡이병의 병 발생을 20.0%까지 현저하게 감소시켰다. DY46 균주의 배양 추출물로부터 8 에서 64 µg mL-1의 농도에서 다양한 식물병원진균의 균사생장 억제 효과를 보여주는 항균 물질인 32,33-didehydroroflamycoin (DDHR)을 동정하였다. DDHR은 1000 μg mL-1의 농도에서 토마토 잿빛곰팡이병을 완벽하게 억제하였다. 또한 1 mL 당 104 포자 농도로 접종한 토마토에서 DDHR (100 μg mL-1)은 토마토 잿빛 곰팡이병의 병 발생을 11.1%까지 감소시켰으며 DDHR을 처리한 토마토에서 DDHR은 짧은 잔류시간 (2일)을 보여주었다. S. plumbeus strain CA5 균주는 춘천 지역에서 수집한 토양으로부터 분리하였다. CA5 균주의 세포 추출물은 포도 잿빛 곰팡이병의 병 발생을 22.2%까지 현저하게 감소시켰다. Lucensomycin은 1 µg mL-1의 농도에서 다양한 식물병원진균의 균사생장을 강하게 억제하는 유효 성분으로 동정되었으며 lucensomycin은 100 μg mL-1의 농도에서 포도 잿빛곰팡이병을 완벽하게 억제하였다. 이러한 결과는 AK 방출 활성을 가지는 미생물 대사산물이 식물병 방제와 특히 강한 효과와 짧은 잔류 활성이 요구되는 저장병 방제를 위한 살균활성 물질의 풍부한 저장소가 될 수 있다는 것을 보여준다.

Deep Learning Networks and ICT-based Plant Disease and Animal Activity Detection System for Digital Agriculture

부젤아닐 경상국립대학교 대학원 2022 국내박사

The majority of food for human beings comes from agriculture. Recently, farmers have had significant pressure to fulfill the rising demand for agricultural products with the increased world population. However, various factors such as catastrophic diseases, urbanization, and climate changes limit agriculture production. Moreover, conventional and subsistence farming cannot meet the increased global food requirement. In this context, it is of utmost necessity to apply the latest technology and tools in agriculture for food safety and production increment. Therefore, the conventional farming concept has been quickly transitioning into digital farming. The main objective of this study was to implement deep learning networks and information and communication technology (ICT) to detect plant diseases, segment and measure disease severity, and detect animal activity. The varieties of deep convolutional neural networks were applied and evaluated their performances. This study has been divided broadly into two parts. The first part deals with the tomato disease classification using lightweight attention-based convolutional neural networks and strawberry gray mold disease segmentation and severity measurement. Likewise, the second part contains the pig posture and locomotion activity detection system using the deep learning-based object detection models and tracking algorithm. Two experiments were conducted on plant disease classification and segmentation and one on pig posture and walking activity detection. In the first experiment of plant disease identification, ten varieties of tomato diseases and healthy leaves images were collected from both the open-source database and the glasshouse located at Gyeongsang National University. A lightweight attention-based deep convolutional neural network (ACNN) was designed to improve the performance of the model for plant disease classification. The total images were divided into training, testing, and validation datasets at a ratio of 8:1:1. Then the performance of the proposed model was compared with the baseline CNN without attention (WACNN) and the standard ResNet50 model. In the second experiment, three concentrations of Botrytis cinerea (causal agent of gray mold disease) were inoculated to the strawberry plants at an early reproductive stage. The occurrence of disease spots on the leaves and their expansion were recorded using a handheld RGB camera daily non-invasively. The raw images were pre-processed to remove clutter background and to extract the target leaf only. Then a deep CNN-based pixel-level segmentation Unet model was designed, trained, tested, and validated using the pre-processed images. The performance of the deep learning model was calculated using the standard segmentation metrics (pixel accuracy, intersection over union (IoU) accuracy, and dice accuracy) and validated using the 5-fold cross-validation method. Moreover, the performance of the Unet model was compared with the XGboost and K-means machine learning models and an image processing algorithm. The disease severity is calculated by using the percentage of diseased pixels in a leaf. The results of tomato disease classification showed that the deep CNN with attention mechanism improved by 1.2% in the tomato disease classification accuracy compared to CNN without attention mechanism in compensation of a few more network parameters and complexity. The CNN without attention module extracts the global features from the whole image. However, the characteristics of the diseased regions would be more specific to an individual disease class. Therefore, the attention module emphasizes regional features rather than global features. Thus, boosting the disease classification accuracy of the model. Whereas, in terms of gray mold disease segmentation, the average pixel, dice, and IoU accuracies of 98.24%, 89.71%, and 82.12%, respectively, were achieved from the Unit model, followed by XGBoost (98.06%, 87.76%, and 80.12%) on 80 test images. Results showed that the Unit model surpasses the conventional XGBoost, K-means, and image processing technique in detecting and quantifying the gray mold disease. The Unit model has two encoder and decoder blocks without fully connected networks. Thus, the network parameters reduce considerably, allowing the model to converge even in a small number of training datasets. Moreover, the Unet model provided a disease segmented image of the same size as the input image due to implementing an up-converter block. For the pig posture and walking activity detection, an experiment was conducted in the experimental pig barn located in Gyeongsang National University. The concentration of greenhouse gases (GHGs) was elevated by closing the ventilator and door of the pig barn for an hour three times a day (morning, day, and night), and the treatment was repeated for three days. The GHGs concentration before, after, and after an hour of treatment was measured by taking air samples from three spatial locations near the center of the house and analyzed using gas chromatography (GC). The livestock environment monitoring system (LEMS) collected the other environmental data (temperature and humidity), including CO2 gas. A top view network camera (HIK VISION) was installed to record the videos of pig activities and stored them in a network video recorder (NVR). A total of 6,012 frames from the video were labeled manually using the computer vision annotation tool (CVAT) and split into training and testing datasets (9:1). Three variances of object detection models (YOLOv4, Faster R-CNN, and SSD ResNet) were trained and validated to detect pig postures and walking activity. Then the deep association simple online real-time tracking algorithm (Deep SORT) was implemented to track the individual pig in the video clips. Pig postures and walking activity information was extracted from the one hour before, during, and after the treatment periods and analyzed the changes in activity due to the compromised environment. The pigs' standing, walking, and sternal lying activities reduce with increased GHGs, increasing lateral lying posture duration. Also, the pigs were more active in the morning than daytime and the least in the nighttime. Moreover, the pig posture detection performances of the object detection models were evaluated using the average precision (AP) and mean AP (mAP) and found the YOLO model provided the highest mAP of 98.67%, followed by the Faster R-CNN model (96.42%). Furthermore, the YOLO model outperformed in terms of detection speed (0.031 s/frame), followed by the SSD model (0.123 s/frame) and the Faster R-CNN model (0.15 s/frame). Therefore, the deep learning networks showed that they could effectively solve complex agricultural problems. However, more researches are recommended for further improvement. Finally, a web-based client-server architecture (http://sfsl.gnu.ac.kr) was designed to automatically collect the environmental and image data from the experimental sites. Similarly, a multi-user python interactive program called JupyterHub was installed on the server (https://sfslws.gnu.ac.kr), allowing the deep learning models to run in the cloud.

Invasive Exotic Diseases on Dogwood, Boxwood and Impatiens in the United States

Daughtrey, Margery L Cornell University ProQuest Dissertations & Theses 2023 해외박사(DDOD)

Dogwood and boxwood are two mainstays of American gardens: eastern flowering dogwood and Pacific dogwood trees offer the beauty of their large white bracts, and boxwood shrubs offer stately versatile evergreen foliage with resistance to deer browse. The popular bedding plant impatiens has provided spreading masses of color in shaded landscapes for decades, particularly since new hybrids were introduced in the 1960s. Each of these plants has been a beloved ornamental, contributing to societal well-being as well as serving as a source of profits for the nursery-greenhouse industry. Toward the end of the 20th century, native dogwood species on both coasts of North America were threatened by an introduced fungus, while at the beginning of the 21st century boxwood encountered an introduced fungus, and impatiens an introduced oomycete, a downy mildew. The ability of each of these plants to provide beauty in the landscape was destroyed by the long-distance movement of a pathogen from a still unknown geographical location. In the case of the two native dogwoods, prior to the introduction of their new fungal pathogen, the trees were supplying ecosystem contributions in woodlands in addition to having an important role in the ornamentals industry. Although global plant movement is regulated, pathogens are microscopic, often internal, and easily overlooked, especially when introduced on a host of low susceptibility that is relatively symptom-free. This dissertation includes an introductory chapter (Chapter 1) that describes the introduction of three new pathogens into the United States: the fungi Discula destructiva and Calonectria pseudonaviculata and the oomycete Plasmopara destructor. These pathogens have had extraordinary impact on dogwood, boxwood and impatiens, respectively. I contributed to developing a proper understanding of each of these new diseases, working closely with others at botanical gardens, plant breeding companies, USDA, and other universities to conduct the needed research and extension. My contributions have been primarily in the areas of diagnosis and management-in particular, I remain keenly interested in solutions to these problems that will come through ever-improving plant breeding and selection. This first chapter provides a current perspective on these diseases, building upon the stories told in the published articles which constitute Chapters 2, 3 and 4 of the dissertation. Chapter 2 is a co-authored invited review article from 1994, which covers dogwood anthracnose from its first appearance on both American coasts until the review's publication. Craig R. Hibben and I collaborated closely on field, laboratory, and plot studies from the beginning of the discovery of dogwood anthracnose in New York, which was less than 20 years before the article was written. Chapter 3 is an invited review article on boxwood blight prepared more recently, in 2019, in which I covered all aspects of the disease from the literature available and my own experience. Chapter 4 deals with the impatiens downy mildew and is a very different type of publication from the others. It is a synergistic collaboration of five scientists from three states that covered outbreaks of disease in the production industry in 2019 and 2020 that occurred on disease-resistant impatiens. Our goal was to make the flower industry aware that high disease resistance is not sufficient to prevent disease losses when environment and other circumstances strongly favor downy mildew development. Having conducted trials that repeatedly demonstrated the superiority of the new lines of impatiens, it was important to me that the limits of their resistance to downy mildew also be properly understood. In addition to providing more recent information than offered by the earlier published articles that make up Chapters 2, 3 and 4, Chapter 1 includes my interpretation of what we have learned from these three new diseases of ornamental plants. One obvious conclusion is that introduction of new pathogens of ornamentals is not a rare event, so ornamentals plant pathologists should be ever-vigilant.

Isolation and characterization of antifungal metabolites from Maesa japonica, Pterocarya tonkinensis, and Trichoderma longibrachiatum against plant pathogenic fungi : Maesa japonica, Pterocarya tonkinensis 및 Trichoderma longibrachiatum 유래 천연물의 식물병원균에 대한 항균활성

Men Thi Ngo UNIVERSITY OF SCIENCE AND TECHNOLOGY 2022 국내박사

In the course of screening for plants and microbes showing antifungal activity, two plant species Maesa japonica and Pterocarya tonkinensis, and a fungal strain Trichoderma longibrachiatum SFC10166 exhibited promising in vivo antifungal activities against phypathogenic fungi. In this study, 37 compounds were isolated based on a bioassay-guided fractionation. The isolated compounds were five new acylated triterpenoid saponins including maejaposide I (1), maejaposides C-1, C-2, and C-3 (24), and maejaposide A-1 (5) from M. japonica, two new natural compounds including pterocaryalactone (7) and pterocaryafuranone (8) from P. tonkinensis, and two new metabolites including spirosorbicillinol D (25) and 2′,3′-dihydro-epoxysorbicillinol (26) from T. longibrachiatum SFC10166, along with 28 known compounds. These chemical structures were determined by NMR, CD, MS data, and a comparison of their NMR, CD, and MS data with those reported in the literature. Based on the results of the in vitro antifungal assay, Magnaporthe oryzae and Phytophthora infestans were the most sensitive to the isolated compounds among the test phytopathogenic fungi. When the compounds were applied onto plants at a concentration of 500 µg/mL, compounds 27, 14, 21, and 22 effectively suppressed rice blast disease with control values over 85%, while compounds 21, 22, 28, 30, 31, and 35 strongly reduced the development of late blight on tomato plants with control values ranging of 71 to 96%, compared to the untreated control. Taken together, our results suggest that the M. japonica, P. tonkinensis, T. longibrachiatum SFC10166, and their substances can be used as a source to develop natural fungicides. 식물병 방제효과를 나타내는 식물 및 미생물 소재를 탐색하는 과정에서, 강한 살균 활성을 나타내는 2종의 식물 Maesa japonica 및 Pterocarya tonkinensis와 1종의 곰팡이 균주 Trichoderma longibrachiatum SFC10166가 선발되었다. 본 연구에서는 살균활성 기반의 크로마토그래피 기법을 사용하여 총 37개의 화합물을 분리하였다. M. japonica으로부터 5 개의 새로운 사포닌 화합물 maejaposides I, C-1, C-2, C-3 및 A-1 (15); P. tonkinensis에서 2 개의 새로운 천연 화합물 pterocaryalactone (7)과 pterocaryafuranone (8); T. longibrachiatum SFC10166으로부터 2 개의 새로운 천연 화합물spirosorbicillinol D (25)와 2′,3′-dihydro-epoxysorbicillinol (26)을 분리했고, 나머지 28개의 화합물은 알려진 화합물이다. 화학 구조 동정을 위해서nuclear magnetic resonance (NMR), circular dimorphism (CD), mass spectrometry (MS) 분석 결과와 문헌 정보를 비교하여 분리한 화합물의 화학 구조를 동정하였다. In vitro 항균활성 시험 결과, 분리화합물은 시험한 식물병원균 중 Magnaporthe oryzae 와 Phytophthora infestans 에 가장 강한 살균 활성을 나타냈다. 화합물은 500 µg / mL의 농도로 식물에 처리했을 때, 화합물 27, 14, 21 및 22 는 벼 도열병을 무처리구 대비 85% 이상 방제하는 우수한 효과를 보였고 화합물 21, 22, 28, 30, 31 및 35는 토마토 역병을 무처리구 대비 96% 방제하는 우수한 효과를 보였다. 본 실험결과는 M. japonica 추출물, P. tonkinensis 추출물, T. longibrachiatum SFC10166 배양액 및 이로부터 분리한 화합물이 천연 살균제 개발에 이용될 수 있음을 나타내는 흥미로운 결과이다.

Improving Abiotic and Biotic Stress Tolerance in Floriculture Crops

South, Kaylee Anne The Ohio State University ProQuest Dissertations & 2020 해외박사(DDOD)

An intensive production system is used to produce greenhouse floriculture crops, marketed for their flowers and attractive foliage. Chemical, environmental, and cultural methods are used to manage biotic and abiotic stresses during production. Additional tools are needed by growers because of growing concerns around the negative impact of plant production on humans and the environment. The objective of this research was to evaluate potential tools to improve floriculture crop resilience under stress during production and post-production.Botrytis cinerea causes disease in most major greenhouse crops and is resistant to several fungicides. Additional control methods, like plant growth promoting bacteria (PGPB) that can improve plant performance by increasing plant resilience to stress are needed. A collection of 60 bacterial strains was evaluated in a dual culture assay and an initial greenhouse trial with Petunia x hybrida Carpet Red Bright’ to identify strains for the biocontrol of B. cinerea. Daily flower disease severity ratings were used to select seven strains that were evaluated in the validation greenhouse trial. Three Pseudomonas strains were selected for the greatest reduction in B. cinerea infection.The efficacy of PGPB and the plant’s susceptibility to B. cinerea were affected by fertilization. Petunia x hybrida `Carpet Red Bright’ was treated with bacteria or a commercial biocontrol product and fertilized with synthetic chemical or organic fertilizer at a low or high rate. Measured plant growth and flower disease severity revealed that plants with the high rate synthetic fertilizer were the largest and had the lowest disease severity. Reduction of disease severity varied between bacterial and fertilizer treatment combinations. Plants treated with one bacterium had reduced disease severity at the high rate synthetic chemical fertilizer but not at the low rate organic fertilizer.Specific fertility programs provide crops with needed macro and micronutrients, but overuse can lead to negative environmental impacts, plant disorders, and higher susceptibility to other stresses. Application of PGPB can improve floriculture plant performance grown under low fertility. Ninety-four bacterial isolates identified from the rhizosphere of ornamental plants were evaluated in P. hybrida `Picobella Blue’ grown under low fertility, and 15 isolates were selected for increasing plant performance. Whole-genome sequencing was used to determine their identity and bacteria were evaluated again under low fertility along with untreated plants receiving higher fertilizer rates. Three bacteria were selected as top performers for increases in flowering, vegetative health, and vegetative quality.Over- or under-watering is a common stress for Phalaenopsis orchids once the orchid reaches the consumer, but to avoid water stress, ice cube irrigation is recommended. Orchid health was evaluated after irrigation with either ice cubes or room temperature water. Orchids grown in bark media did not have a reduction in display life or health when irrigated with ice cubes.Utilizing novel tools, like PGPB or ice cube irrigation, during production or post-production is an important move toward improving floriculture crop performance. These tools can be used to improve floriculture crop resilience to biotic or abiotic stresses and establish additional sustainable practices for the greenhouse industry.

A Study on Efficient Data Augmentation Method for Automatic Classification of Plant Disease

이새봄 가천대학교 일반대학원 2023 국내석사

본 논문은 이미지로부터 식물 병해충의 패턴을 효과적으로 추출할 수 있는 데이터 증강 방법을 제안한다. 병해충에 감염된 식물은 갈변, 점무늬, 가는 실을 꼰 형태 등 다양한 패턴이 존재한다. 식물 병해충을 분류하는 딥러닝 모델의 성능을 올리기 위해서는 병해충의 패턴을 효과 적으로 추출할 수 있는 데이터 증강 방법이 필요하다. 본 연구에서는 60,000장 이상의 병해충 이미지를 수집하여 빅데이터를 구축하고 24가 지 종류의 식물 병해 이미지에 적합한 데이터 증강 방법을 분석한다. 데이터 증강 방법의 분석 결과는 병해충의 진단을 위한 딥러닝 기반 병 해충 자동 분류 네트워크의 데이터 전처리 방법으로 사용한다. 네트워 크를 구성하는 모델은 모두 사전 훈련한(pre-trained) VGGNet, ResNet, DenseNet, EfficientNet, ViT, DeiT이다. 실험 결과 6개의 모델은 97% 이상의 accuracy와 f1 score를 도출했다. 그러므로 본 연구에서 제안하는 데이터 증강 방법은 병해충 분류 문제에 우수한 성능을 보인다. This paper proposed a data augmentation method to extract plant disease patterns from images effectively. Plants infected with diseases have various patterns, such as browning, dot pattern, and thin braided yarn. In order to improve the performance of deep learning models that classify plant diseases, a data augmentation method that can effectively extract pest patterns is needed. In this study, we collected over 60,000 disease images to build big data, and data augmentation methods suitable for 24 crop pest images were analyzed. The analysis result of the data augmentation method is used as a pre-processing data method of deep learning-based automatic disease classification network for diagnosis. The network configuration models are all pre-trained VGGNet, ResNet, DenseNet, EfficientNet, ViT, and DeiT. As a result of the experiment, six models derived accuracy and an f1 score of over 97%. Therefore, the method proposed in this study showed excellent performance in classifying plant diseases.

Characterization and antifungal activity of streptavidins and cationic peptides isolated from streptomyces species effective on fusarium wilt control

전병준 Graduate School, Korea University 2021 국내박사

Characterization and Antifungal Activity of Streptavidins and Cationic Peptides Isolated from Streptomyces Species Effective on Fusarium Wilt Control Fusarium wilt is a devastating soil-borne disease resulting in losses and decreasing the quality and safety of agricultural crops. In this study, macromolecules screening which focuses on the antifungal activity against Fusarium oxysporum f.sp. lycopersici was performed using a natural product library comprising microbial metabolites. Streptomyces cinnamonensis strain KPP02129 and Streptomyces spororaveus strain KPP03845 were found to harbor potent antifungal proteins and peptides in their culture filtrate. Two biotin-binding proteins, streptavidin C1 and streptavidin C2, were identified from culture filtrate of strain KPP02129. The streptavidin C1 has a C-terminal region different from those of known streptavidins which stretches into the biotin-binding site of its monomer. The minimum inhibitory concentration of streptavidin C1 and C2 was found to be 16-64 µg ml−1 for inhibiting the mycelial growth of F. oxysporum f.sp. cucumerinum and F. oxysporum f.sp. lycopersici. Antimicrobial peptide, P3845, was purified from culture filtrate of strain KPP03845. The purified peptide P3845 showed potent minimum inhibitory concentrations (MICs) against various plant pathogenic fungi and bacteria at concentrations 0.125-16 µg mL-1 and 1-32 µg mL-1, respectively. The culture filtrate of the strain KPP02129 and strain KPP03845 markedly reduced disease incidence of Fusarium wilt on tomato plants, showing the potential of macromolecules, streptavidin C1, C2 and P3845 as antifungal agent for the control of plant disease.

Drug Delivery Nanosystems as Plant "Vaccines": Fabrication and Assessment of their use for Plant Protection Against Broad Host-Range Necrotrophic Pathogens

Vega-Vasquez, Pablo Purdue University ProQuest Dissertations & Theses 2020 해외박사(DDOD)

Drug-delivery nano-systems enhances the potency of bioactive molecules due to its increase membrane permeability, as a result of their sub-cellular size. The concept of engineered nanocarriers may be a promising route to address confounding challenges in agriculture that could lead to an increase in crop production while reducing the environmental impact associated with crop protection and food production. A key motivation of this work is to evaluate the potential use of drug delivery nanosystems in agriculture, especially in the area of disease control. To this end, identifying the most suitable materials to serve as carrier and cargo is imperative. Understanding their bioactive properties and their physical-chemical characteristics is critical because these influences not only their biological effects on plants and environmental impact, but also, the fabrication process and potential scaling-up, enabling practical and relevant field applications in the future.In this work, chitosan was selected as nano-carrier material because of its biological and chemical properties. The chemical structure of chitosan allows spontaneous assemble of core-shell like nanostructures via ionic gelation, has enabled it to be used as nano-carrier biomaterial intended for delivery of bioactive cargo. In agriculture, the use of chitosan is of special interest due to its immune-modulatory activity elicited in plants. However, due to its inherent molecular heterogenicity, the formulation and fabrication of stable and low inter-batch variability chitosan nanocarriers via ionic gelation is difficult and time consuming.A myriad of different bioactive molecules has been tested as payload, encapsulated into chitosanbased delivery nano-systems for a range of purposes ranging from biomedicine, pharmaceutical, food and agriculture. In this work plant derived essential oils were selected as bioactive payload. Essential oils are at the core of the plant communication process with their phytobiome, including plant pathogens. Molecules from essential oils can carry an air-borne message serving as a plantto-plant communication system (a phenomenon known as allelopathy) that activate the plant defense mechanisms. Encapsulation of essential oils into chitosan nanocarriers is only possible by forming nano-emulsions.Despite the potential benefits from the use of chitosan and essential oils in agriculture, its use at a large scale has been hindered by the overwhelming inconsistencies in the current literature, regarding their formulation and fabrication. This work addresses these problems and presents evidence that support the feasibility of producing highly chitosan nanocarriers loaded with essential oils, in a facile and rapid way, using FDA-grade materials only, without the need of expensive or specialized instrumentation.The plant-pathogen compatible interaction between A. thaliana and B. cinerea was used as biological model to test the hypothesis that chitosan nano-carriers and essential oil nano-emulsions can enhance the quantitative disease resistance of plants against broad host-range necrotrophic pathogens. We found that these treatments display a dose-dependent response in plants triggering a systemic immune response. Image-based phenotyping analysis showed that chitosan nanoparticles alone, as well as loaded with d-limonene, significantly enhanced the disease resistance of A. thaliana against B. cinerea. Nano-emulsions using essential oils from cinnamon, clove, coriander and red thyme also produced similar effects on the defense response in the pathosystem under study. Functional analysis of the differentially expressed genes from treated plants revealed that these treatments up-regulated the biological process involved in "stress management", while down-regulated the biological process required for normal growth and development during ideal, non-stressful conditions.

The soil borne bacteria, Streptomyces sp. regulates cucumber growth and anti-fungal activity against Sclerotinia sclerotiorum through enzymatic and phytohormonal modulation : 근권 미생물 Streptomyces sp. LH4를 이용한 오이의 균핵병(Sclerotinia sclerotiorum) 방제효과

이원희 경북대학교 대학원 2018 국내석사

In the soil ecosystem, microbial diversity is exist and these diverse organisms are interacting with plant roots and influencing on physicochemical properties of plant. From these diverse of microorganism, some of them are causing disease such as club root by Plasmodiophora brassicae, damping off by Rhizoctonia solani, and soft rot by Erwinia spp. While some are beneficially interacting with plant. Rhizobacteria, bacteria which colonizing the root zone (rhizosphere) of plants is one of well-known beneficial microorganism and it named as plant growth-promoting rhizobacteria (PGPR) which contribute to the promotion of plant growth either direct or indirect ways. This study was conducted to investigate the protease and cellulase medium test by promoting plant growth and isolating bacteria that antagonize the pathogens, and biologically investigated the pathogenicity against pathogenic fungi. Isolated strain LH4 was identified as Streptomyces sp. by 16S rRNA sequencing and phylogenetic analysis. The antifungal antagonistic activity against Sclerotinia sclerotiorum was confirmed by measuring glucanase production and activity of Streptomyces sp. LH4. The treatment of Streptomyces sp. LH4 pure culture medium and Sclerotinia sclerotiorum with cucumber plants inhibited the growth of Streptomyces sp. LH4. In addition, pathogen defense major hormones JA and SA analysis showed that the mutation of two hormones increased the disease resistance of cucumber. As a result of this study, Streptomyces sp. LH4 showed anticancer effects against plant growth promotion and pathogen. 병원성 곰팡이는 식물에게 피해를 주는 요소 중 하나이며, 이를 해결하기 위해 천연추출물, 생물농약 등을 이용한 친환경 방제에 대한 연구가 많이 진행되고 있다. 본 연구는 바닷가 인근 토양에서 분리한 방선균의 식물 생육 촉진과 균핵병 (Sclerotinia sclerotiorum)에 대한 길항능을 확인하기 위해 진행하였다. 분리된 균주 LH4의 phosphate solubilizing, protease, cellulase 활성과 병원성 곰팡이에 대한 항균성을 대치 배양을 통해 조사하였다. LH4는 16S rRNA 염기서열 및 계통발생 분석을 통해 Streptomyces sp.로 동정되었다. Streptomyces sp. LH4의 cellulase 활성도를 fluorescence로 측정한 결과, cellulase의 활성이 지속적으로 증가하는 것을 확인하였다. 또한, Streptomyces sp. LH4 배양액과 Sclerotinia sclerotiorum을 오이 유묘에 관주 처리하였을 때, Streptomyces sp. LH4처리구에서 균핵병 발병이 억제되었고 생육이 촉진되는 것을 확인하였다. 식물의 대표적인 병원 저항성 유도 호르몬인 JA와 SA 분석 결과, 두 호르몬의 상호작용을 통한 LH4균주의 오이 균핵병에 대한 저항성 증대가 확인되었다. 따라서 본 연구 결과를 통해 Streptomyces sp. LH4가 오이 균핵병을 효과적으로 방제하는 친환경 소재로서 활용가능성을 확인하였다.