Deep learning for weed identification based on seed images

Abstract

Reliable plant species identification from seeds is intrinsically difficult due to the scarcity of features and because it requires specialized expertise that is becoming increasingly rarer, as the number of field plant taxonomists is diminishing (Bacher 2012, Haas and Häuser 2005). On the other hand, seed identification is relevant in some science domains such as plant community ecology, archaeology, paleoclimatology. Besides, economic activities such as agriculture, require seed identification to assess weed species contained in the "soil seed banks" (Colbach 2014) to enable targeted treatments before they become a problem. In this work, we explore and evaluate several approaches by using different training image sets with various requisites and assessing their performance with test datasets of different sources. The core training dataset is provided by the Anthos project (Castroviejo et al. 2017) as a subset of its image collection. It consists of nearly a 1000 images of seeds identified by experts. As identification algorithm, we will use state-of-the-art convolutional neural networks for image classification (He et al. 2016). The framework is fully written in Python using the TensorFlow (Abadi et al. 2016) module for deep learning.

A Cloud-Based Framework for Machine Learning Workloads and Applications

“…This model [73], [74] was trained using a dataset from the Spanish Royal Botanical Garden which consists of around 28000 images from 743 species and 493 genera. This classifier can prove to be helpful for farmers looking to classify their seeds or users looking to improve their botanical skills.…”

A new model based on improved VGG16 for corn weed identification

“…As innovations in deep learning theory and hardware conditions continue to develop, people can construct deeper network models to extract more features (Ye et al, 2019;Wagle and Harikrishnan, 2021), and an increasing number of network models are being constructed for use in various aspects of agricultural production (Chakraborty et al, 2021). Deep learning has been widely applied in recent years, especially in smart agriculture fields, such as pest and disease detection (Mique and Palaoag, 2018;Liu et al, 2022;Wu et al, 2022), plant and fruit recognition (Jaiganesh et al, 2020;Bongulwar Deepali, 2021), and crop and weed detection and classification (Pando et al, 2018;Jin et al, 2021). Image recognition technology has long been used for weed recognition applications (Jiang et al, 2020).…”