Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
169 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
45 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Autonomous Sweet Pepper Harvesting for Protected Cropping Systems (1706.02023v1)

Published 7 Jun 2017 in cs.RO

Abstract: In this letter, we present a new robotic harvester (Harvey) that can autonomously harvest sweet pepper in protected cropping environments. Our approach combines effective vision algorithms with a novel end-effector design to enable successful harvesting of sweet peppers. Initial field trials in protected cropping environments, with two cultivar, demonstrate the efficacy of this approach achieving a 46% success rate for unmodified crop, and 58% for modified crop. Furthermore, for the more favourable cultivar we were also able to detach 90% of sweet peppers, indicating that improvements in the grasping success rate would result in greatly improved harvesting performance.

Citations (184)

Summary

  • The paper presents a novel robotic system, Harvey, that automates sweet pepper harvesting in protected cropping environments to reduce manual labor dependency.
  • It employs advanced RGB-D vision algorithms integrated with a 7DOF manipulator and a unique end-effector that decouples gripping and cutting actions.
  • Field experiments showed a 58% overall harvest success and 92% detachment success for the Claire cultivar, highlighting potential for sustainable horticulture.

Autonomous Sweet Pepper Harvesting for Protected Cropping Systems

The paper "Autonomous Sweet Pepper Harvesting for Protected Cropping Systems" by Lehnert et al. investigates an approach to automate the harvesting of sweet peppers using a robotic harvester system named Harvey. The authors identify the significant potential of robotic harvesting to alleviate the reliance of the horticulture industry on manual labor, especially given the high labor costs and market volatility. This research focuses on the development and testing of a system tailored to work within protected cropping environments, which offer a controlled setting that mitigates some environmental challenges such as lighting variability and occlusions.

The authors categorize the main challenges of autonomous harvesting into three domains: detection, grasp selection, and manipulation of the crop. These elements are crucial for the effective detachment and collection of horticultural crops like sweet peppers, which are typically challenging due to their variety of shapes, colors, and orientations. In responding to these challenges, the authors present a modular approach that integrates advanced vision algorithms and a uniquely designed end-effector to optimize the harvesting process.

System Design

The system presented comprises a mobile platform equipped with a 7DOF manipulator and a customized harvesting tool. The end-effector is a critical component featuring a suction gripper and an oscillating blade linked together via a flexible tether, facilitating a sequential process for gripping and cutting the peppers. This decoupling mechanism allows the harvesting tool to adjust dynamically, significantly increasing the success rate of detached sweet peppers from the plant.

Experimental Methodology and Results

Field experiments were conducted using different cultivars in a controlled environment. Notably, the system achieved a 58% harvest success rate in trials with the Claire cultivar, with a remarkable detachment success rate of 92%. These results are an indicator of the system's efficacy, particularly in grasping and detaching peppers without causing damage to the plant or the fruit.

The vision-based algorithm leverages RGB-D data to perform sweet pepper segmentation, effectively addressing challenges related to occlusions and variable light conditions that are endemic in agricultural environments. The approach utilizes a combination of color information and 3D spatial data to ascertain optimal grasp poses, with candidate grasp paths evaluated based on their expected success rate.

Implications and Future Work

The findings contribute to the growing body of research in agricultural automation, presenting solutions with significant implications for the economic and operational models of horticultural production. Improvements in robotic harvesting could lead to more sustainable practices by reducing operational costs, ensuring optimal harvesting cycles, and maintaining high produce quality.

Despite the promising results, there are avenues for refinement and further development. Enhancements to grasp detection, specifically around irregularly shaped peppers and variance in peduncle orientation, will continue to be a priority. Future iterations of this system may also explore advanced perception techniques, such as machine learning-based approaches, for more robust crop and peduncle detection.

The research underscores a step toward more sophisticated, commercially viable autonomous harvesting systems. By addressing both theoretical and practical challenges, the authors provide a foundation that could encourage broader adoption and adaptation of robotic solutions within the agricultural industry.