Fecha de publicación: 08/04/2022 Fuente: Wipo "precision farming" A system for yield estimation of mango fruit crop is disclosed in the present invention where temporal mango fruit dataset created with the captured images of the mango tree is trained by the pre-trained neural network, classifying the input test image of the mango fruit by the pre-trained deep neural network with trained images of the dataset, detecting and labelling the mangoes by a deep learning object detector at all the stages of growth present in the image which help for early yield estimation of mango fruit crop.