Identification of Cowpea Genotypes by Machine Learning Using Digital Images of Pods and Green Beans

Cowpea is a crop of great importance worldwide. Consuming pods and green beans provides vitamins, minerals, and functional components for people with limited access to vegetables. The objective of this study was to use machine learning and artificial intelligence techniques to classify and distinguish cowpea lines based on images of their pods and green beans. Digital images of four heirloom landraces of the cowpea genotypes (Sempre Verde, Rabo de Tatu, Corujinha, and Paulistinha) and nine cultivars (BRS Novaera, BRS Olhonegro, BRS Verdejante, BRS Exuberante, BRS Pajeú, BRS Miranda, IPA 206, BRS Tapaihum, and BRS Pingo de Ouro) were processed using four deep learning architectures for feature extraction (vectorization): InceptionV3, SqueezeNet, VGG16, and VGG19. Six machine learning algorithms were evaluated: K-Nearest Neighbors (KNN), Decision Tree, Random Forest (RF), Gradient Boosting (GB), Support Vector Machines (SVM), and Multi-Layer Perceptron (MLP). The MLP (Multi-Layer Perceptron) and SVM models performed best, particularly when integrated with the InceptionV3 embedder. In terms of classification efficiency, the evaluated algorithms achieved exceptional accuracy, reaching a Classification Accuracy (CA) of 0.999 for pod discrimination and 1.000 for green beans. For pod classification, these models achieved high performance metrics, including an Area Under the Curve (AUC) of 1.000. For green beans, the MLP maintained a high Area Under the Curve (AUC = 1.000) and better probabilistic calibration (lower Log-Loss) than the SVM. Digital image-based identification using machine learning is an efficient, non-destructive approach for morphological characterization and discrimination of cowpea genotypes, supporting high-throughput phenotyping applications.