Scientists tap AI to analyze citrus rind color for optimized harvest schedules

04 Dec 2023 --- Researchers in China have developed an AI-based framework for predicting and visualizing citrus fruit color transformation in orchards, which indicates when fruits are mature enough for harvest. The technique is slated for enhancing fruit development and optimal harvest timing, with potential applications in other citrus species and fruit crops.

The team has also created an Android application that processes citrus images and a specified time interval, generating a future color image of the fruit.

As per the FAO, world citrus production and export have grown steadily over the past three decades, with China, Brazil and the US leading their production.

However, the crop is facing slow production challenges, underlines the study, which is published in Plant Phenomics. 

The research was conducted at the Huazhong Agricultural University, Wuhan, China and emphasizes the need for F&B to focus on improving fruit quality and post-harvest processes.

“Citrus rind color is a good indicator of fruit development and methods to monitor and predict color transformation help the decisions of crop management practices and harvest schedules,” note the authors.

Fruit maturity and harvest
According to the authors, color change is a crucial indicator of fruit maturity, which has traditionally been gauged by human judgment.

For most citrus species, the green color of the rind diminishes because of the presence of pigments and other colors, such as yellow, orange and red, emerge when carotenoids accumulate.

“Because this process is strongly correlated to fruit development, citrus color is a good indicator of the maturity stage, and its proper monitoring enables better decisions regarding crop management practices. Moreover, the optimal flavor and storage life can only be obtained if the fruit is harvested at a specific maturity stage,” the authors state. 

Thus, accurate color prediction in such fruits is essential to promote fruit quality and arrangement of the harvest schedule.

For the study, the researchers observed 107 Navel orange images during color transformation to validate and train the network. The framework utilizes a deep mask-guided generative network for accurate predictions and has a design requiring fewer resources that allow mobile device implementation.

The authors created the dataset by field investigation and image acquisition (Image credit: Plant Phenomics).

According to the researchers, the technique can replicate color transformation accurately, even with different viewing angles and the color of oranges.

Sensory analysis of the samples validated the network’s effectiveness, with a majority finding high similarity between synthesized and real images.

Challenges with previous techniques
Recent machine vision and neural network advancements offer more objective and robust color analysis, but they “struggle with varying conditions and translating color data into practical maturity assessments,” states the study.

The current techniques to predict color transformation in fruits have some limitations, note the researchers.

The Asian agriculture sector is leveraging AI power to manage labor-intensive, high-tech farming tasks.“Research gaps remain in predicting color transformation over time and developing user-friendly visualization techniques.” Further, it is challenging to put these methods to use in agriculture as they have limited computing capabilities. This drives the demand for optimized, efficient technologies to analyze and predict color changes to gauge maturity.

Additionally, the framework’s adaptability to edge devices like smartphones makes it highly “practical for in-field use,” which signifies its potential in agriculture and other fields in the future. 

AI in agriculture
The Asian agriculture sector is witnessing a digital transformation, marked by the growing adoption of AI to manage labor-intensive, high-tech farming tasks. Such advancements are changing how traditional Asian farms cultivate and harvest, offering solutions to sustainable farming and rural development.

AI is also making inroads into mushroom cultivation farms. In October, Tupu introduced a decentralized farming system that combines its modular farming system (patent-pending) with bioscience, robotics and AI to grow organic gourmet mushrooms “directly” in cities.

Meanwhile, ReSeed is using its AI-powered digital ledger transparency platform to collect and process data for carbon credit measurement protocols to incentivize farmers to use sustainable practices in the field.

These developments align with Givaudan and Canvas8’s recent report that entails “Augmented Assistance” as one of the five macro trends set to transform the F&B industry.

By Insha Naureen