Purpose

The Agricultural Image Repository (AgIR) serves as a resource for advancing computer vision, artificial intelligence, and image-based phenotyping within precision agricultural. It aims to support the development of automated tools for weed management, crop monitoring, and plant trait analysis. By providing a rich and diverse collection of labeled images representing various weeds, cover crops, and cash crops under diverse conditions, AgIR empowers researchers and developers to create innovative solutions for sustainable and efficient agricultural practices.

Scope

AgIR is made-up of 3 broad plant types; weeds, cash crops, and cover crops, each of which can be further divided into two main scene types, Semi-Field and true real-world Field each providing a unique set of advantages and use cases.

Plant Types

AgIR encompasses a broad range of agricultural imagery, specifically focusing on:

• Weeds: A diverse collection of weed species, over 40 agronomically relevant species across different regions in the United States.

• Cover Crops: Common cover crop

• Cash Crops: Key cash crop species grown in the US.

• Scene Types: Images capture both semi-field and real-world field settings, providing a comprehensive view of plant growth environments.

Scene Types

AgIR encompasses two distinct scene types, each with unique advantages and applications.

Semi-Field Data

• Environment: Images captured in semi-controlled environments (e.g., nurseries) using the BenchBot, a gantry-like robotic system.

• Throughput: High-throughput collection allows for the acquisition of large amounts of data daily (over 500 images across a large nursery potting area) across 3 US locations.

• Annotation: Facilitated by the plain black background of weed fabric and hand weeds, enabling automatic annotation through photogrammetry, digital image processing, and deep learning.

• Trade-offs: While offering scalability and automation, these images may not fully represent real-world plant conditions.

...

Field Data

• Environment: Images collected manually by teams in real-world field settings.

• Real-World Relevance: Images closely reflect natural plant growth and environmental variations.

• Annotation: Images captured with a grey mat background to aid in segmentation and annotation processes.

• Trade-offs: Limited by manual collection, resulting in a smaller data pool compared to semi-field data.

Complementary Nature

The semi-field and field data work in tandem to overcome their individual limitations. The extensive, automatically annotated semi-field data can be leveraged to train annotation helper models, which then aid in the annotation of the more limited, real-world field data. This approach addresses one of the most challenging and expensive tasks in deep learning model development: image labeling.

Target Users

• Researchers: Scientists in computer vision, artificial intelligence, and agricultural disciplines seeking data to develop and refine algorithms for weed management, crop monitoring, and image-based phenotyping.

• Developers: Software engineers building applications for automated weed detection, plant trait analysis, and precision agriculture tools.

• Agronomists: Agricultural professionals interested in utilizing computer vision and AI-powered tools for decision-making in weed management and crop production.

• Industry Professionals: Representatives from agricultural companies, technology providers, and input suppliers seeking data-driven solutions for improving agricultural practices.

• Crop Groups: Organizations focused on specific crops, that invest in research and development to enhance crop production and sustainability.

An Agricultural Image Repository for the United States

For this repository to be comprehensive and include most plant species of agricultural relevance for the different regions in the US, partnering with institutions throughout the country was key. There are nine states and ten institutions participating in data collection for the AgIR in 2023 (Figure 1). These partnerships allow us to gather images of the plant species growing in different environments, hence capturing the variability that it can introduce in morphology, and facilitates image collection particularly of weeds which can be challenging to grow at will.

...

Plant Species of Agricultural Relevance for the US

The list of target plant species for 2022 and 2023 includes the most common cover crop and cash crop species for the US. Moreover, the weed species list was defined by the participating partners to whom we asked to provide the names of the 4 most relevant weed species in their regions. The current list of species includes 44 weed species, 15 cover crop species and 3 cash crop species.

...