ABSTRACT The ranking of seed lots is a fundamental process for all companies in the seed industry. This work aims to demonstrate data mining methods for ranking sorghum seed lots during the seed processing through analysis of quality control data. Germination and cold tests were performed to verify the physiological quality of the lots. Seed samples from each lot were evaluated in two moments: post-cleaning and finished product (ready for marketing). The results after pre-processing totaled 188 rows of data with six attributes, encompassing 150 lots accepted for marketing, 6 rejected, and 32 intermediate lots. The classifiers used were J48, Random Forest, Classification Via Regression, Naive Bayes, Multilayer Perceptron, and IBk. The Resample filter was used for adjustment of the data. The k-fold technique was used for training, with ten folds. The metrics of Accuracy, Precision, Recall, F-measure, and ROC Area were used to verify the accuracy of the algorithms. The results obtained were used to determine the best machine-learning algorithm. IBk and J48 presented the highest accuracy of data; the IBk technique presented the best results. The Resample filter was essential for solving the data imbalance problem. Sorghum seed lots can be classified with great accuracy and precision through artificial intelligence and machine learning technique.

RESUMO A classificação de lotes de sementes é um processo fundamental para todas as empresas do setor sementeiro. O objetivo do trabalho é demonstrar os métodos de mineração de dados de ranqueamento de lotes de sementes de sorgo durante o processo de beneficiamento, através de análises de dados do controle de qualidade. Os testes realizados foram germinação e teste de frio, com o objetivo de verificar a qualidade fisiológica dos lotes. As amostras de sementes de cada lote foram avaliadas em dois momentos: póslimpeza e produto acabado (pronto para comercialização). Os dados gerados, após o pré-processamento, totalizaram 188 linhas com seis atributos, contabilizando 150 lotes aceitos para comercialização, seis rejeitados e 32 denominados intermediários. Os classificadores utilizados foram J48, Random Forest, Classification Via Regression, Naive Bayes, Multilayer Perceptron e IBk. Utilizou-se o filtro Resample para ajustamento dos dados. A técnica empregada para treinamento foi a k-fold, com 10 folds. Para verificar a precisão dos algoritmos foram utilizadas as métricas de Acurácia, Precisão, Recall, F-measure e Área ROC. Com os resultados obtidos determinou-se o melhor algoritmo de aprendizagem de máquina. Verificou-se que o IBk e o J48 obtiveram maior acurácia nos dados, sendo que a técnica de IBk obteve o melhor resultado. O filtro Resample foi importante para resolver o problema do desequilíbrio dos dados. Concluímos ser possível classificar lotes de sementes de sorgo com grande acurácia e precisão através de inteligência artificial e sua técnica de aprendizado de máquina.

ABSTRACT The widespread adoption of sensor technology has made it a standard practice for obtaining precise and timely information during the harvest and post-harvest periods. One sensor that has gained popularity for post-harvest seed monitoring is the MQ-3, which identifies ethanol in the air as products undergo fermentation. However, these sensors typically require a stable operation. This study aimed to assess the stabilization time of an MQ-3 sensor when measuring ethanol levels in anaerobic bean seeds. We used six bean seed samples, each with an average moisture content of around 14%. We employed a completely randomized experimental design with nine repetitions for each sample. Every repetition consisted of 25 bean seeds placed in sealed flasks containing 70 mL of distilled water. This setup induced anoxic conditions within the flask, promoting anaerobic respiration in the seeds. After 24 hours, we exposed an air sample to the MQ-3 sensor and took readings at various time intervals (12-14, 19-21, 36-38, 68-70, 130-132, 192-194, 314-316, 616-618 seconds). The average stabilization time for the MQ-3 sensor while quantifying ethanol concentrations in the bean samples were approximately 23 seconds. The sensor demonstrated efficacy, convenience, and rapidity in assessing ethanol levels in anaerobic bean seeds.

ABSTRACT To enhance speed and agility in interpreting physiological quality tests of seeds, The use of algorithms has emerged. This study aimed to identify suitable machine learning models to assist in the precise management of seed lot quality. Soybean lots from two companies were assessed using the Supplied Test Set, Cross-Validation (with 8, 10, and 12 folds), and Percentage Split (with 66% and 70%) methods. Variables analyzed through Tetrazolium tests included vigor, viability, mechanical damage, moisture damage, bed bug damage, and water content. Method performance was determined by Kappa, Precision, and ROC Area metrics. Classification Via Regression and J48 algorithms were employed. The technique utilizing 66% of data for training achieved 93.55% accuracy, with Precision and ROC Area reaching 94.50% for the J48 algorithm. Applying the cross-validation method with 10 folds resulted in 90.22% of correctly classified instances, with a ROC Area outcome like the previous method. Tetrazolium Vigor was the primary attribute used. However, these results are specific to this study's database, and careful planning is necessary to select the most effective application methods.

RESUMO No setor de produção de sementes, é imprescindível o uso de equipamentos rápidos e práticos que auxiliem nos processos de tomada de decisão. Assim, o uso de sensores que podem capturar informações para auxiliar na agricultura tem contribuído significativamente para o cultivo da soja em todo o mundo, sendo utilizado em diversas etapas, seja em pré ou pós-colheita. Objetivou-se desenvolver um protótipo de um etilômetro com sensor MQ-3 e avaliar seu tempo de estabilização quando utilizado em lotes de sementes de soja submetidos à hipóxia. A pesquisa foi realizada com 10 lotes de sementes de soja, oriundas de cinco cultivares (NS 8383 RR, SOY Peso RR, M 8372 IPRO, M8644 IPRO e M8808 IPRO), a qual foi dividida em duas etapas, a primeira correspondendo ao desenvolvimento do equipamento de aferição do etanol com o sensor MQ-3 e a segunda relacionada ao tempo de estabilização do sensor (3, 6, 9, 12, 15, 18, 21, 24, 27 e 30 s) em sementes de soja submetidas ao processo de anaerobiose. O equipamento, bem como o sensor MQ-3 apresentaram estabilidade de aferição média aos 16 s após seu acionamento, sendo recomendado esse tempo para o uso do etilômetro em sementes de soja. Por meio desse mesmo sensor MQ-3, é possível, com eficiência, a detecção de etanol presente no ar liberado por sementes de soja durante o processo de anaerobiose.

ABSTRACT In the seed production sector, it is essential to use fast and practical equipment to help decision-making processes. Thus, the use of sensors that can capture information to assist in agriculture has contributed significantly to soybean cultivation worldwide, being used in several stages, whether pre-harvest or post-harvest. Thus, this study aimed to develop a prototype of ethylometer with an MQ-3 sensor and to evaluate its stabilization time when used in soybean seed lots submitted to hypoxia. The research was carried out with 10 lots of soybean seeds from five cultivars (NS 8383 RR, SOY Peso RR, M 8372 IPRO, M8644 IPRO and M8808 IPRO) and was divided into two stages. The first one corresponds to the development of the ethanol measuring equipment with an MQ-3 sensor, and the second one to measurement of the sensor’s stabilization time (3, 6, 9, 12, 15, 18, 21, 24, 27, and 30 s) in soybean seeds submitted to anaerobiosis. The equipment and the MQ-3 sensor showed the stability of average measurement at 16 s after activation. This time is recommended for use of the ethylometer in soybean seeds. Through the MQ-3 sensor, it is possible to efficiently detect ethanol present in the air released by soy seeds during the anaerobic process.

ABSTRACT The seed germination and vigor evaluation are essential for the sowing sector to measure the performance of different seed lots and improve the efficiency of storage and sowing processes. However, the analysis of various tests to determine seed quality generates a large amount of information, making it almost impossible for humans to perform a quick and effective quality control analysis. Therefore, the objective of this study was to evaluate the differences in the physiological quality of soybean seeds in different cultivars using machine learning techniques to rank the lots based on their quality. Three cultivars were used, and the analysis was germination, accelerated aging, tetrazolium treatment, seedling emergence, and 1000 seed weight from 65 lots were measured. The lots were evaluated in two phases, one immediately after harvest and the other after six months of storage. Random forest, multi-layer perceptron, J48, and classification via regression classifiers were used, aided by the feature resampler technique. Random forest and classification via regression obtained the highest accuracy, and the random forest technique obtained the best results. Therefore, it is possible to classify soybean seed lots with great accuracy and precision using artificial intelligence and machine learning techniques.

ABSTRACT The seed sector faces several challenges when it comes to ensuring a quick and accurate decision making when working with large amounts of data on physiological quality of seed lots, which makes the process time-consuming and inefficient. Thus, artificial intelligence (AI) emerges as a new technological option in the seed sector to solve database problems in the post-harvest stages. This study aims to use machine learning to classify maize seed lots. Data were obtained from eight maize seed crops from a private company. These data were mined using the following classifiers: J48 (DecisionTree), RandomForest, CVR (ClassificationViaRegression ) , lBk (lazy.IBK), MLP (MultiLayerPercepton), and NäiveBayes. Cross-validation was used for data measurement, with the data set, including training and testing data, being divided into 10 subsets. The described steps were performed using the Weka software. It is concluded that results obtained allow the classification of maize seed lots with high accuracy and precision, and these algorithms can better classify the maize seed lot through vigor attributes, thus enabling more accurate decision making based on vigor tests on a reduced evaluation time.

ABSTRACT New, efficient, low-cost techniques for image processing and alternative machine learning for seed processing are of academic and industrial interest. This study aims to identify fissures in bark and peeled rice seeds using X-ray and RGB image processing techniques and machine learning. Samples of three batches of rice seeds were used: a batch of seeds not subjected to drying (peeled seed), and the other two comprised of dried seeds, one containing seeds with husk and another containing huskless seeds; each sample comprised 100 seeds. Images in X-ray and RGB formats were provided in the sequence processed in ImageJ software and introduced in the machine learning software, where they were pre-processed using the appropriate filters and then classified by the J48 and linear discriminant analysis (LDA) classifiers. X-ray images obtained using differentiated equipment allow the identification of cracks in rice seeds using image processing techniques and the LDA classifier. Capturing images using RGB is a viable alternative. Using filters, either individually or in combination, may constitute an adequate alternative for rice seed classification.