Julio Barón-Velandia

Precision agriculture has increasingly adopted controlled agricultural microclimates, particularly smart greenhouses, as a strategy to enhance crop yields while maintaining environmental conditions within suitable ranges for each crop. Among the variables that govern the water balance in these systems, air humidity plays a critical role; therefore, accurate humidity forecasting is essential for implementing timely control actions that support productivity levels. However, greenhouse conditions are frequently perturbed by extreme weather events, which lead to nonlinear and non-stationary humidity dynamics. In this context, the aim of this study was to design an optimized evolving fuzzy inference system for humidity forecasting that can adapt to changing and unforeseen situations in agricultural microclimates. A prototyping-based methodology was followed, including phases of communication, quick planning, modeling and quick design, construction of the prototype, and deployment. A hybrid genetic algorithm was used to optimize the parameters of an evolving Mamdani-type fuzzy inference system, extended to handle missing values in online data streams. Thirty independent optimization runs were performed, and the best configuration achieved a mean squared error of 1.20 × 10⁻² in humidity forecasting using one minute of data for three months. The resulting model showed high interpretability, with an average number of 1.35 rules, tolerance for missing values, imputing 2% of the data, and robustness to sudden changes in the data stream with a p-value of 0.01 for the Augmented Dickey–Fuller test at alpha = 0.05. In general, the optimized evolving fuzzy inference system obtained an effectiveness rate greater than 90% and demonstrated adaptability to extreme weather conditions, suggesting its applicability to other phenomena with similar characteristics.

The bacterium Xanthomonas campestris poses a significant threat to global agriculture due to its ability to infect leaves, fruits, and stems under various climatic conditions. Its rapid spread across large crop areas results in economic losses, compromises agricultural productivity, increases management costs, and threatens food security, especially in small-scale agricultural systems. To address this issue, this study developed a model that combines fuzzy logic and neural networks, optimized with intelligent algorithms, to detect symptoms of this foliar disease in 15 essential crop species under different environmental conditions using images. For this purpose, Sugeno-type fuzzy inference systems and adaptive neuro-fuzzy inference systems (ANFIS) were employed, configured with rules and clustering methods designed to address cases where diagnostic uncertainty arises due to the imprecision of different agricultural scenarios. The model achieved an accuracy of 93.81%, demonstrating robustness against variations in lighting, shadows, and capture angles, and proving effective in identifying patterns associated with the disease at early stages, enabling rapid and reliable diagnoses. This advancement represents a significant contribution to the automated detection of plant diseases, providing an accessible tool that enhances agricultural productivity and promotes sustainable practices in crop care.

Greenhouse cultivation is one of the current strategies to address the challenges of food production, sustainability, and food quality. Similarly, the use of technological tools to automate greenhouse environments through a set of sensors and actuators allows for the control and improvement of processes within this environment. This document presents data collected from the sensors and actuators of two identical greenhouse environments, one with the cultivation of stringless blue lake beans and the other without cultivation. The aim is that this dataset will provide a broader characterization of the behavior of climatic variables inside greenhouse environments and how they are impacted by control actions, subsequently contributing to the development of new research on implementations of or improvements to control, supervision, management, and automation actions in greenhouse environments.

Currently, the increase in devices capable of continuously collecting data on non-stationary and dynamic variables affects predictive models, particularly if they are not equipped with algorithms capable of adapting their parameters and structure, causing them to be unable to perceive certain time-varying properties or the presence of missing data in data streams. A constantly developing solution to such problems is evolving fuzzy inference systems. The aim of this work was to systematically review forecasting models implemented through evolving fuzzy inference systems, identifying the most common structures, implementation outcomes, and predicted variables to establish an overview of the current state of this technique and its possible applications in other unexplored fields. This research followed the PRISMA methodology of systematic reviews, including scientific articles and patents from three academic databases, one of which offers free access. This was achieved through an identification, selection, and inclusion workflow, obtaining 323 records on which analyses were carried out based on the proposed review questions. In total, 62 investigations were identified, proposing 115 different system structures, mainly focused on increasing precision, in addition to addressing eight main fields of application and some optimization techniques. It was observed that these systems have been successfully implemented in forecasting variables with dynamic behavior and handling missing values, continuous data flows, and non-stationary characteristics. Thus, their use can be extended to phenomena with these properties.

Obtaining highly accurate predictive models to precisely estimate corn crop yields is essential for making informed decisions in a sustainable agriculture environment. There are various approaches to achieve this goal, including models based on Fuzzy Logic, Association Rules, and Machine Learning. However, some of these models have limitations in terms of the accuracy of their predictions, attributable to the high complexity and non-linearity in the interactions between factors. While Machine Learning techniques alone can achieve high precision, the inclusion of multiple attributes can reduce it. This study focuses on identifying the most influential factors at the regional level through a comprehensive analysis of the relevance of features associated with corn crop yields in Colombia, a country in the Neotropical zone. To accomplish this, climatological time series and historical yield records are used through a methodology based on CRISP- DM, widely used in the field of data mining, involving a review of related work, data cleaning and transformation, relevance evaluation using the RReliefF algorithm, and verification of the performance of the most influential factors through prediction algorithms. The results obtained demonstrate that solar radiation, precipitation, vapor pressure, and maximum and minimum temperatures exert the greatest influence on corn crop yield prediction, with a relevance factor of 0.033, 0.032, 0.026, 0.022, and 0.021, respectively. In the validation of the performance of the selected factors, two predictive models were implemented. The first, based on Artificial Neural Networks, yielded a RMSE of 0.1216 with the subset of variables and 0.1403 with all available variables. In the second, Linear Regression was applied, resulting in an RMSE of 0.1417 with the subset of variables and 0.1424 with all available variables. These results highlight the importance of the selected features as the most influential climatic factors in defining highly accurate predictive models in the Neotropical zone.

This chapter shows the outcomes for four optimization models based on fuzzy inference systems, intervened using Quasi-Newton and genetic algorithms, to early assess bean plants’ leaves for Xanthomonas campestris disease. Plant diseases are an important threat to food production. While major pathogenicity determinants required for disease have been extensively studied, less is known on how pathogens thrive during host colonization, especially at early infection stages. This research is classified as experimental and it is focused on the provided dataset analysis to establish the critical variables, and the most relevant tools for pattern and behavior recognition. The RGB colour intensity for the data sets and photographs used to analyse the model implementation defines the classification of the plant's condition (healthy or insanity). The best model performance is 99.68% when compared with the training data and a 94% effectiveness rate on the detection of Xanthomonas campestris in a bean leave image. One of the most important characteristics of neural networks is their high accuracy to the cost of interpretability, nevertheless, the best model for this research developed using a fuzzy inference system does not sacrifice interpretability. Therefore, these results would allow farmers to take early measures to reduce the impact of the disease on the look and performance of green bean crops. The model has high accuracy and interpretability when optimized and a greater capacity to detect the existence of the disease in a plant.

Establishing the indoor and outdoor humidity values in a greenhouse allows us to describe the crop yield during its entire developmental cycle. This study seeks to develop a predictive model of indoor relative humidity values in a greenhouse with high accuracy and interpretability through the use of optimized fuzzy inference systems, in order to offer greenhouse users a clear and simple description of their behaviour. The three-phase methodology applied made use of descriptive statistics techniques, correlation analysis, and prototyping paradigm for the iterative and incremental development of the predictive model, validated through error measurement. The research resulted in six models which define the behaviour of humidity as a result of temperature, CO₂, and soil moisture, with percentages of effectiveness above 90. The implementation of a Mamdani-type fuzzy inference system, optimized by a hybrid method combining genetic and interior point algorithms, allowed to predict the relative humidity in greenhouses with high interpretability and precision, with an effectiveness percentage of 90.97 and MSE (mean square error) of 8.2e3.

The degree of roasting of the coffee determines the physical properties of the bean which are directly represented in the quality of the coffee, to classify the coffee bean efficiently represents a challenge that has been addressed from different technological approaches with colorimeters. This research aims to simplify the identification of the roast level of ground coffee on the Agtron scale by characterizing the degree of roast using an optoelectronic color sensor and establishing a correlation between the Red, Green, and Blue (RGB) scales. This allows for the assurance of quality levels of the beans right from the roasting process. This research comprehends the collection and preparation of samples, the definition of RGB and CIE L*a*b* values, and their interpretation in the Agtron scale using the red component of the RGB scale. The results showed an efficient and accurate estimation for the roast degree of ground coffee beans (0.1371 MSE) that uses minimum processing requirements and a function to assess the intermediate values in the Agtron scale. The characterization of the roast degree of ground coffee beans using data collected from an optoelectronic color sensor through a high-precision function with a linear structure enables the description of intermediate values not fully represented on the Agtron scale. This enhances the process of identifying the roast degree, facilitating subsequent quality assurance processes by maintaining the beans at the desired roast level. Key words: Agtron; coffee bean color; color sensor; RGB.