PREVISÃO DE MEDIDA DE PESO BOVINO UTILIZANDO REDES NEURAIS CONVOLUCIONAIS

This study presents a solution for predicting weight measurements in cattle through machine learning techniques using images. Weighing cattle is crucial for optimizing production and ensuring animal welfare, but the traditional method can cause stress to the animal. The aim is to develop a model based on convolutional neural networks to estimate the weight of cattle through image analysis, minimizing the need for direct handling. The main contributions include the proposal of a convolutional neural network model, the comparison of different image filters (RGB and Grayscale), and the analysis of various perspectives of the cattle (lateral and frontal). The expected results seek to validate the efficiency of the proposed model and contribute to more sustainable and efficient practices in livestock farming.