Comparison of Different Classifiers and the Majority Voting Rule for the Detection of Plum Fruits in Garden Conditions.

Razieh Pourdarbani Sajad Sabzi Mario Hernández Hernández José Luis Hernández-Hernández Ginés García_Mateos Davood Kalantari José Miguel Molina Martínez (2019, [Artículo])

Color segmentation is one of the most thoroughly studied problems in agricultural applications of remote image capture systems, since it is the key step in several different tasks, such as crop harvesting, site specific spraying, and targeted disease control under natural light. This paper studies and compares five methods to segment plum fruit images under ambient conditions at 12 different light intensities, and an ensemble method combining them. In these methods, several color features in different color spaces are first extracted for each pixel, and then the most e

ective features are selected using a hybrid approach of artificial neural networks and the cultural algorithm (ANN-CA). The features selected among the 38 defined channels were the b* channel of L*a*b*, and the color purity index, C*, from L*C*h. Next, fruit/background segmentation is performed using five classifiers: artificial neural network-imperialist competitive algorithm (ANN-ICA); hybrid artificial neural network-harmony search (ANN-HS); support vector machines (SVM); k nearest neighbors (kNN); and linear discriminant analysis (LDA). In the ensemble method, the final class for each pixel is determined using the majority voting method. The experiments showed that the correct classification rate for the majority voting method excluding LDA was 98.59%, outperforming the results of the constituent methods.

remote sensing in agriculture artificial neural network hybridization environmental conditions majority voting plum segmentation INGENIERÍA Y TECNOLOGÍA CIENCIAS TECNOLÓGICAS TECNOLOGÍA DE LOS ALIMENTOS

Procesos oceanográficos asociados al desarrollo y dispersión de florecimientos de Gymnodinium catenatum en el norte del Golfo de California

Oceanographic processes associated with harmful algal blooms of Gymnodinium catenatum development and dispersal in the northern Gulf of California

GABRIELA RESENDIZ COLORADO (2023, [Tesis de doctorado])

Durante los últimos años, durante la temporada de invierno y primavera, se ha detectado la presencia de florecimientos algales nocivos (FAN) causados por el dinoflagelado Gymnodinium catenatum en el norte del golfo de California (NGC). Estos eventos tienen impactos ecológicos, económicos y sociales negativos porque G. catenatum es una especie productora de saxitoxina, la cual, está asociada al envenenamiento paralítico por consumo de mariscos, lo que origina que la autoridad sanitaria se vea en la necesidad de implementar vedas en áreas de extracción de almeja generosa en el NGC al detectar producto contaminado. Por lo tanto, es necesario conocer los procesos físicos que provocan la ocurrencia y recurrencia de estos eventos, así como contar con un sistema de monitoreo y alerta temprana que permita tomar decisiones y acciones oportunas de mitigación contra los efectos perjudiciales de estos fenómenos. En este trabajo se abordaron estas necesidades de investigación por medio de la implementación de un método de detección remota de estos FAN, utilizando la clasificación de máxima verosimilitud basada en los datos de dos eventos sucedidos en 2015 y 2017. Los resultados de este enfoque fueron satisfactorios al reproducir la temporalidad de la presencia de la especie documentada por medio de muestreos semanales en la bahía de San Felipe, así como de la detección geográfica en las áreas que se conocen que son afectadas recurrentemente. Para estudiar los procesos físicos asociados a estos FAN, se implementó un modelo hidrodinámico usando el Sistema de Modelación Oceánica Regional (ROMS) para el norte del golfo de California. A partir de este modelo se obtuvieron datos de variables como energía cinética turbulenta, temperatura, corrientes y se complementaron con datos de esfuerzo de fondo producido por oleaje estimados a partir de datos de la quinta generación del reanálisis atmosférico del clima global del ECMWF (ERA-5), los análisis de estas variables y la abundancia semanal de G. catenatum permitieron identificar que los principales procesos asociados a la formación de estos florecimientos es el incremento del esfuerzo de fondo asociado al oleaje y la turbulencia. Estos, a su vez, tienen una relación con el cambio en el patrón del viento que se caracteriza por ser del noroeste durante las temporadas de invierno y primavera. Respecto a la dispersión de los florecimientos algales, con base en los datos obtenidos del modelo hidrodinámico, se realizaron experimentos

During the recent winter and spring seasons, harmful algal blooms (HABs) caused by the dinoflagellate Gymnodinium catenatum have been detected in the northern Gulf of California (NGC). These events have negative ecological, economic, and social impacts because G. catenatum is a species producer of saxitoxin, which is associated with paralytic shellfish poisoning, which causes the need to implement bans by the health authority in extraction areas of generous clam in the NGC when detecting contaminated products. Therefore, it is necessary to identify the physical processes that cause the occurrence and recurrence of these events and have a monitoring and early warning system that allows timely decisions and mitigation actions to be taken against the harmful effects of these henomena. In this work, these research gaps are approached by implementing a remote detection method for these HABs, using maximum likelihood classification based on data from two events in 2015 and 2017. The results of this approach were satisfactory by reproducing the temporality presence of Gymnodinium catenatum documented through weekly sampling in San Felipe Bay, as well as geographic detection in areas known to be recurrently affected. To study the physical processes associated with these HABs, a hydrodynamic model was implemented using the Regional Ocean Modeling System (ROMS) for the northern Gulf of California. From this model, variables such as turbulent kinetic energy, temperature, and currents were obtained and were complemented with data on bottom stress produced by waves estimated from the fifth generation ECMWF atmospheric reanalysis of the global climate (ERA-5) data. Analyzing these variables and the weekly abundance of G. catenatum, it was possible that the processes associated with forming these blooms are the increase in bottom stress related to waves and turbulence. These processes are related to the change in the wind pattern characterized by northwest winds during the winter and spring seasons. Regarding the dispersion of algal blooms, based on the data obtained from the hydrodynamic model, lagrangian experiments were carried out to estimate the transport of the HAB. The results showed that the modeled dispersion corresponds with the detection carried out with the remote sensing method mplemented in this work for the 2017 event. The results obtained from this work are essential knowledge for the operational implementation of monitoring and early warning systems ..

Gymnodinium catenatum, florecimientos algales nocivos, percepción remota, modelación hidrodinámica, norte del golfo de California : Gymnodinium catenatum, harmful algal bloom, remote sensing, hydrodynamic modeling, northern Gulf of California BIOLOGÍA Y QUÍMICA CIENCIAS DE LA VIDA OTRAS ESPECIALIDADES DE LA BIOLOGÍA OTRAS OTRAS