Master of Science / Department of Mechanical and Nuclear Engineering / Dale Schinstock / Developments in farm related technology have increased the importance of mapping individual plants in the field. An automated mapping system allows the size of these fields to scale up without being hindered by time-intensive, manual surveying. This research focuses on the development of a mapping system which uses geo-located images of the field to automatically locate plants and determine their coordinates. Additionally, this mapping process is capable of differentiating between groupings of plants by using Quick Response (QR) codes. This research applies to green plants that have been grown into seedlings before being planted, known as transplants, and for fields that are planted in nominally straight rows. The development of this mapping system is presented in two stages. First is the design of a robotic platform equipped with a Real Time Kinematic (RTK) receiver that is capable of traversing the field to capture images. Second is the post-processing pipeline which converts the images into a field map. This mapping system was applied to a field at the Land Institute containing approximately 25,000 transplants. The results show the mapped plant locations are accurate to within a few inches, and the use of QR codes is effective for identifying plant groups. These results demonstrate this system is successful in mapping large fields. However, the high overall complexity makes the system restrictive for smaller fields where a simpler solution may be preferable.

<p>Six methods are used for clustering data containing two different objects: sugar-beet plants </p><p>and weed. These objects are described by 19 different features, i.e. shape and color features. </p><p>There is also information about the distance between sugar-beet plants that is used for </p><p>labeling clusters. The methods that are evaluated: k-means, k-medoids, hierarchical clustering, </p><p>competitive learning, self-organizing maps and fuzzy c-means. After using the methods on </p><p>plant data, clusters are formed. The clusters are labeled with three different proposed </p><p>methods: expert, database and context method. Expert method is using a human for giving </p><p>initial cluster centers that are labeled. The database method is using a database as an expert </p><p>that provides initial cluster centers. The context method is using information about the </p><p>environment, which is the distance between sugar-beet plants, for labeling the clusters. </p><p> </p><p>The algorithms that were tested, with the lowest achieved corresponding error, are: k-means </p><p>(3.3%), k-medoids (3.8%), hierarchical clustering (5.3%), competitive learning (6.8%), self- </p><p>organizing maps (4.9%) and fuzzy c-means (7.9%). Three different datasets were used and the </p><p>lowest error on dataset0 is 3.3%, compared to supervised learning methods where it is 3%. </p><p>For dataset1 the error is 18.7% and for dataset2 it is 5.8%. Compared to supervised methods, </p><p>the error on dataset1 is 11% and for dataset2 it is 5.1%. The high error rate on dataset1 is due </p><p>to the samples are not very well separated in different clusters. The features from dataset1 are </p><p>extracted from lower resolution on images than the other datasets, and another difference </p><p>between the datasets are the sugar-beet plants that are in different growth stages. </p><p> </p><p>The performance of the three methods for labeling clusters is: expert method (6.8% as the </p><p>lowest error achieved), database method (3.7%) and context method (6.8%). These results </p><p>show the clustering results by competitive learning where the real error is 6.8%. </p><p> </p><p>Unsupervised-learning methods for clustering can very well be used for plant identification. </p><p>Because the samples are not classified, an automatic labeling technique must be used if plants </p><p>are to be identified. The three proposed techniques can be used for automatic labeling of </p><p>plants.</p>

Six methods are used for clustering data containing two different objects: sugar-beet plants and weed. These objects are described by 19 different features, i.e. shape and color features. There is also information about the distance between sugar-beet plants that is used for labeling clusters. The methods that are evaluated: k-means, k-medoids, hierarchical clustering, competitive learning, self-organizing maps and fuzzy c-means. After using the methods on plant data, clusters are formed. The clusters are labeled with three different proposed methods: expert, database and context method. Expert method is using a human for giving initial cluster centers that are labeled. The database method is using a database as an expert that provides initial cluster centers. The context method is using information about the environment, which is the distance between sugar-beet plants, for labeling the clusters. The algorithms that were tested, with the lowest achieved corresponding error, are: k-means (3.3%), k-medoids (3.8%), hierarchical clustering (5.3%), competitive learning (6.8%), self- organizing maps (4.9%) and fuzzy c-means (7.9%). Three different datasets were used and the lowest error on dataset0 is 3.3%, compared to supervised learning methods where it is 3%. For dataset1 the error is 18.7% and for dataset2 it is 5.8%. Compared to supervised methods, the error on dataset1 is 11% and for dataset2 it is 5.1%. The high error rate on dataset1 is due to the samples are not very well separated in different clusters. The features from dataset1 are extracted from lower resolution on images than the other datasets, and another difference between the datasets are the sugar-beet plants that are in different growth stages. The performance of the three methods for labeling clusters is: expert method (6.8% as the lowest error achieved), database method (3.7%) and context method (6.8%). These results show the clustering results by competitive learning where the real error is 6.8%. Unsupervised-learning methods for clustering can very well be used for plant identification. Because the samples are not classified, an automatic labeling technique must be used if plants are to be identified. The three proposed techniques can be used for automatic labeling of plants.

In today’s increasingly fast-moving, complex, and competitive world, the need for flexibility and creativity in teaching and learning is crucial. For that reason, innovative educational methods should be introduced. In education, web-based learning and portable devices are emerging as teaching and learning aids which can be efficient and effective tools. Learning use and identification of ornamental plants are the main objectives of the plant materials courses offered by Department of Plant and Soils Sciences at Mississippi State University (MSU). The professors, teaching assistants (TA), and students use the MSU gardens to study and identify ornamental plant species. This can be time consuming for both instructors and students. This research developed an automated web-based database system to deliver information on the ornamental plants in the MSU gardens. Apache, MySQL, PHP, JavaScript, Dreamweaver, and Photoshop software were used to develop this application in the Windows environment and information about each plant was entered into the database. Plant locations were given by longitude and latitude coordinates and linked to Google maps. Quick Response codes(QR code) were created to directly access ornamental plant information at the field. This database may function as a virtual TA for the plant materials courses and as an information source for the public. Users can search the ornamental plant information and determine the location of plants using a computer or mobile device. Plant information can be retrieved from the field by a smart phone with a QR code reader. To evaluate the effectiveness and efficiency of developed automated system, an experimental study and questionnaire survey were designed.

Redes complexas é uma área de estudo relativamente recente, que tem chamado a atenção da comunidade científica e vem sendo aplicada com êxito em diferentes áreas de atuação tais como redes de computadores, sociologia, medicina, física, matemática entre outras. Entretanto a literatura demonstra que poucos são os trabalhos que empregam redes complexas na extração de características de imagens para posterior analise ou classificação. Dada uma imagem é possível modela-la como uma rede, extrair características topológicas e, utilizando-se dessas medidas, construir o classificador desejado. Esse trabalho objetiva, portanto, investigar mais a fundo esse tipo de aplicação, analisando novas formas de modelar uma imagem como uma rede complexa e investigar diferentes características topológicas na caracterização de imagens. Como forma de analisar o potencial das técnicas desenvolvidas, selecionamos um grande desafio na área de visão computacional: identificação vegetal por meio de análise foliar. A identificação vegetal é uma importante tarefa em vários campos de pesquisa como biodiversidade, ecologia, botânica, farmacologia entre outros. / Complex networks is a relatively recent field of study, that has called the attention of the scientific community and has been successfully applied in different areas such as computer networking, sociology, medicine, physics, mathematics and others. However the literature shows that there are few works that employ complex networks in feature extraction of images for later analysis or classification. Given an image, it can be modeled as a network, extract topological features and, using these measures, build the classifier desired. This work aims, therefore, investigate this type of application, analyzing new forms of modeling an image as a complex network and investigate some topological features to characterize images. In order to analyze the potential of the techniques developed, we selected a major challenge in the field of computer vision: plant identification by leaf analysis. The plant identification is an important task in many research fields such as biodiversity, ecology, botany, pharmacology and others.

Redes complexas é uma área de estudo relativamente recente, que tem chamado a atenção da comunidade científica e vem sendo aplicada com êxito em diferentes áreas de atuação tais como redes de computadores, sociologia, medicina, física, matemática entre outras. Entretanto a literatura demonstra que poucos são os trabalhos que empregam redes complexas na extração de características de imagens para posterior analise ou classificação. Dada uma imagem é possível modela-la como uma rede, extrair características topológicas e, utilizando-se dessas medidas, construir o classificador desejado. Esse trabalho objetiva, portanto, investigar mais a fundo esse tipo de aplicação, analisando novas formas de modelar uma imagem como uma rede complexa e investigar diferentes características topológicas na caracterização de imagens. Como forma de analisar o potencial das técnicas desenvolvidas, selecionamos um grande desafio na área de visão computacional: identificação vegetal por meio de análise foliar. A identificação vegetal é uma importante tarefa em vários campos de pesquisa como biodiversidade, ecologia, botânica, farmacologia entre outros. / Complex networks is a relatively recent field of study, that has called the attention of the scientific community and has been successfully applied in different areas such as computer networking, sociology, medicine, physics, mathematics and others. However the literature shows that there are few works that employ complex networks in feature extraction of images for later analysis or classification. Given an image, it can be modeled as a network, extract topological features and, using these measures, build the classifier desired. This work aims, therefore, investigate this type of application, analyzing new forms of modeling an image as a complex network and investigate some topological features to characterize images. In order to analyze the potential of the techniques developed, we selected a major challenge in the field of computer vision: plant identification by leaf analysis. The plant identification is an important task in many research fields such as biodiversity, ecology, botany, pharmacology and others.

Nous étudions la problématique de classification dite fine en se concentrant sur la détermination des espèces botaniques à partir d’images de feuilles. Nous nous intéressons aussi bien à la description et la représentation de l’objet qu’aux algorithmes de classification et des scénarios d’identification utiles à l’utilisateur. Nous nous inspirons du processus manuel des botanistes pour introduire une nouvelle représentation hiérarchique des feuilles. Nous proposons aussi un nouveau mécanisme permettant d’attirer l’attention au tour de certains points caractéristiques de l’objet et d’apprendre des signatures spécifiques à chaque catégorie.Nous adoptons une stratégie de classification hiérarchique utilisant une série de classifieurs locaux allant des plus grossiers vers les plus fins; la classification locale étant basée sur des rapports de vraisemblance. L’algorithme fournit une liste d’estimations ordonnées selon leurs rapports de vraisemblance. Motivés par les applications, nous introduisons un autre scénario proposant à l’utilisateur un ensemble de confiance contenant la bonne espèce avec une probabilité très élevée. Un nouveau critère de performance est donc considéré : la taille de l’ensemble retourné. Nous proposons un modèle probabiliste permettant de produire de tels ensembles de confiance. Toutes les méthodes sont illustrées sur plusieurs bases de feuilles ainsi que des comparaisons avec les méthodes existantes. / We introduce models for fine-grained categorization, focusing on determining botanical species from leaf images. Images with both uniform and cluttered background are considered and several identification scenarios are presented, including different levels of human participation. Both feature extraction and classification algorithms are investigated. We first leverage domain knowledge from botany to build a hierarchical representation of leaves based on IdKeys, which encode invariable characteristics, and refer to geometric properties (i.e., landmarks) and groups of species (e.g., taxonomic categories). The main idea is to sequentially refine the object description and thus narrow down the set of candidates during the identification task. We also introduce vantage feature frames as a more generic object representation and a mechanism for focusing attention around several vantage points (where to look) and learning dedicated features (what to compute). Based on an underlying coarse-to-fine hierarchy, categorization then proceeds from coarse-grained to fine-grained using local classifiers which are based on likelihood ratios. Motivated by applications, we also introduce on a new approach and performance criterion: report a subset of species whose expected size is minimized subject to containing the true species with high probability. The approach is model-based and outputs a confidence set in analogy with confidence intervals in classical statistics. All methods are illustrated on multiple leaf datasets with comparisons to existing methods.

Made available in DSpace on 2017-07-21T14:19:27Z (GMT). No. of bitstreams: 1 Voncarlos Marcelo Araujo.pdf: 3791024 bytes, checksum: c5d2b6c030643b2e46f5ae7004f73ca8 (MD5) Previous issue date: 2016-03-04 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / The biodiversity of plant species plays a key role in the Earth's ecology, providing food, shelter and maintaining a healthy breathable atmosphere for all living beings. The plants also have medicinal properties and are used for alternative energy sources, such as biofuel. However, the number of plants endangered has gradually increased and the difficulties in the plants manual recognition process, does become a complex and slow task. A viable method for the identification of plants, or to provide a categorization of the plant, is the plant image acquisition and use pattern recognition techniques. In this way, the use of computers, despite having little contribution in the area, can provide important information on the taxonomy of plants, and can serve as a basis for systems that perform tasks such as the selection of certain plants or to guide the specialist for possible decision-making. This paper proposes a method for classification of plants based on collaborative images of the world experts. This method is able to deal with some complexities imposed during the capture of images, as the presence of noise (lighting, shadows and undesirable objects) and plants position variations. To accomplish this task are used texture descriptors based on SIFT, SURF and HOG, which have shown excellent results in several works. To enable testing of the proposed method, we used an image provided by the global task basis for recognition of plants in 2011, ImageCLEF, containing about 2,586 plant samples composed by 41 species divided into two distinct categories: the first one with 13 species and images with presence of noise, and with the second species and 28 sheets of images plotted on a white background. The results of the experiments show that the classifiers trained with texture descriptors are able to achieve good hit rates close to 70%, given the complexity of the problem. Classifiers combination methods have also been used and have been shown capable to improve the performance of classifiers, especially in the test with images that has the presence of noises. / A biodiversidade das espécies de plantas desempenha um papel fundamental na ecologia da Terra, fornecendo alimento, abrigo e mantendo uma atmosfera respirável saudável para todos os seres vivos. As plantas também têm propriedades medicinais e são utilizadas para fontes alternativas de energia, como o biocombustível. No entanto, o número de plantas em risco de extinção tem aumentado gradativamente e as dificuldades presentes no processo manual de reconhecimento de plantas, torna esta tarefa muito complexa e morosa. Uma metodologia viável para a identificação das plantas, ou para fornecer uma categorização de plantas, é a aquisição da imagem da planta e o uso técnicas de reconhecimento de padrões. Dessa forma, o uso da computação, apesar de ainda ter pequena contribuição na área, pode prover informações importantes sobre a taxonomia das plantas, além de poder servir como base para sistemas que executem tarefas como a de seleção de determinado tipo de plantas ou que guiem o especialista para possíveis tomadas de decisões. Neste trabalho é proposto um método para classificação de plantas baseado em imagens colaborativas de especialistas do mundo inteiro. Esse método é capaz de lidar com algumas complexidades impostas durante a captura das imagens, como a presença de ruídos (luminosidade, sombras e objetos indesejáveis) e variações de posições das plantas. Para cumprir essa tarefa são utilizados descritores de textura baseados em SIFT, SURF e HOG, que têm mostrado excelentes resultados em diversos trabalhos. Para possibilitar os testes do método proposto, foi empregada uma base de imagens disponibilizada pela tarefa mundial de reconhecimento de plantas em 2011, ImageCLEF, que contém cerca de 2.586 amostras de plantas composta por 41 espécies divididas em duas categorias distintas: a primeira com 13 espécies e imagens com presença de ruídos, e a segunda com 28 espécies e imagens de folhas plotadas em um fundo branco. Os resultados dos experimentos mostram que os classificadores treinados com descritores de textura são capazes de atingir boas taxas de acertos, próximas a 70%, dada a complexidade do problema. Métodos de combinação de classificadores também foram utilizados e se mostraram capazes de melhorar o desempenho dos classificadores, principalmente nos testes com imagens que tem a presença de ruídos.

Submitted by Rosivalda Pereira (mrs.pereira@ufma.br) on 2017-06-23T20:42:44Z No. of bitstreams: 1 CesarCasteloBranco.pdf: 11257769 bytes, checksum: 911c33f2f0ba5c1c0948888e713724f6 (MD5) / Made available in DSpace on 2017-06-23T20:42:44Z (GMT). No. of bitstreams: 1 CesarCasteloBranco.pdf: 11257769 bytes, checksum: 911c33f2f0ba5c1c0948888e713724f6 (MD5) Previous issue date: 2016-12-12 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPQ) / This work proposes new methodologies to optimize the choice of the parameters of the proportionate normalized least-mean-square (PNLMS) adaptive algorithms. The proposed approaches use procedures based on two optimization methods, namely, the golden section and tabu search methods. Such procedures are applied to determine the optimal parameters in each iteration of the adaptation process of the PNLMS and improved PNLMS (IPNLMS) algorithms. The objective function for the proposed procedures is based on the a posteriori estimation error. Performance studies carried out to evaluate the impact of the PNLMS and IPNLMS parameters in the behavior of these algorithms shows that, with the aid of optimization techniques to choose properly such parameters, the performance of these algorithms may be improved in terms of convergence speed for the identification of plants with high sparseness degree. The main goal of the proposed methodologies is to improve the distribution of the adaptation energy between the coefficients of the PNLMS and IPNLMS algorithms, using parameter values that lead to the minimal estimation error of each iteration of the adaptation process. Numerical tests performed (considering various scenarios in which the plant impulse response is sparse) show that the proposed methodologies achieve convergence speeds faster than the PNLMS and IPNLMS algorithms, and other algorithms of the PNLMS class, such as the sparseness controlled IPNLMS (SC-IPNLMS) algorithm. / Neste trabalho, novas metodologias para otimizar a escolha dos parâmetros dos algoritmos adaptativos LMS normalizados proporcionais (PNLMS) são propostas. As abordagens propostas usam procedimentos baseados em dois métodos de otimização, a saber, os métodos da razão áurea e da busca tabu. Tais procedimentos são empregados para determinar os parâmetros ótimos em cada iteração do processo de adaptação dos algoritmos PNLMS e PNLMS melhorado (IPNLMS). A função objetivo adotada pelos procedimentos propostos é baseada no erro de estimação a posteriori. O estudo de desempenho realizado para avaliar o impacto dos parâmetros dos algoritmos PNLMS e IPNLMS no comportamento dos mesmos mostram que, com o auxílio de técnicas de otimização para escolher adequadamente tais parâmetros, o desempenho destes algoritmos pode ser melhorado, em termos de velocidade de convergência, para a identificação de plantas com elevado grau de esparsidade. O principal objetivo das metodologias propostas é melhorar a distribuição da energia de ativação entre os coeficientes dos algoritmos PNLMS e IPNLMS, usando valores de parâmetros que levam ao erro de estimação mínimo em cada iteração do processo de adaptação. Testes numéricos realizados (considerando diversos cenários nos quais a resposta impulsiva da planta é esparsa) mostram que as metodologias propostas alcançam velocidades de convergência superiores às dos algoritmos PNLMS e IPNLMS, além de outros algoritmos da classe PNLMS, tais como o algoritmo IPNLMS com controle de esparsidade (SCIPNLMS).