Articles

Overview of Button Mushroom Harvesting Technologies

Harvest of cultivated button mushrooms (Agaricus bisporus) is a labor-intensive process due to the delicate nature of the crop and growing conditions. This article surveys current automation technologies and strategies to reduce labor requirements.

Download Save for later Print

Updated:

August 20, 2025

Skip to the end of the images gallery

Robotic Mushroom Harvester

Skip to the beginning of the images gallery

Introduction

Agaricus bisporus (button mushroom) is a highly nutritious edible mushroom with recognized medicinal value. The United States ranks among the top global producers, with total market sales reaching 659 million pounds and a value of $1.09 billion in the 2023-2024 season. Pennsylvania leads the national production, accounting for nearly two-thirds of the mushrooms consumed in the U.S. Of the total output, 88% was destined for the fresh market (USDA-NASS, 2024), which traditionally relies on manual harvesting to maintain excellent quality, shelf-life, and visual appeal (Zied & Arturo, 2017). However, mushroom production in the US has declined by over 30% from 2014 to 2024 (USDA-NASS, 2018, 2021, 2024). The reason is largely attributed to labor shortages, a growing concern across the agricultural sector (Salzwedel, 2023). According to the USDA Economic Research Service, the availability of agricultural labor continues to diminish, posing a serious threat to sustainable mushroom production (ERS, 2023). Addressing this challenge requires the development and adoption of innovative mechanical and robotic harvesting solutions to reduce reliance on manual labor.

Mechanical Mushroom Harvesters

Mechanical mushroom harvesting machines were developed by companies such as Van den Top (Barnveld, the Netherlands) to improve the speed and efficiency of picking high-quality mushrooms compared to manual methods. It features a large front-mounted brush that lifts mushrooms upright after cutting, while a wire mesh removes most of the soil. An adjustable knife provides controlled cutting, and an integrated conveyor system transports the mushrooms and prevents them from falling. However, the machine requires specially designed aluminum mushroom shelves. Additionally, it harvests mushrooms nonselectively, regardless of their size or maturity, often resulting in damage, making the harvested product more suitable for processing rather than fresh consumption.

Robotic Mushroom Harvesting

A robotic mushroom harvesting system consists of several key components, including a machine vision system for mushroom detection and maturity assessment, a picking strategy, a picking mechanism, and an integrated system to approach and interact with the targeted mushrooms.

Mushroom Detection using Machine Vision Systems

With advances in artificial intelligence, machine vision has become increasingly applicable in the mushroom industry. By collecting RGB-D image datasets from mushroom tubs (Figure 1A) and training deep learning models, highly accurate systems for mushroom detection and localization are developed (Figure 1B). These systems enable the identification of mushroom size, clustering patterns, and spatial arrangements, providing essential data for targeted harvesting.

Two phtos showing Mushroom imaging from the side and top of the tub. — **Figure 1.** **(A)** Image dataset preparation by capturing RGB-D images from a mushroom tub, **(B)** Detection output showing identified button mushrooms using a trained deep learning model. Credit: Sadjad Mahnan

Mushroom Maturity Detection Using Machine Vision Systems

Accurate localization and sizing of mushrooms enable the detection of their maturity. Maturity is typically determined by the degree of cap openness, which can be quantified through cap curvature. This curvature is derived from 3D information obtained via stereo vision. Depth data, combined with trained deep learning models, allows for the classification of mushrooms according to their growing stages.

Mushroom Picking Strategy

Due to their clustered growth and varying cap sizes and maturity rates, mushrooms often reach maturity at different times and require selective harvesting across multiple picking breaks. This presents a challenge for robotic systems. To address this, it is essential to develop strategies that target only mature mushrooms. Studies have shown that harvesting involves a combination of bending, twisting, and lifting motions. Therefore, implementing effective picking sequences and adapting the bending direction for detachment are critical for successful robotic harvesting systems (Figure 2)

Photo 1 shows mushroom tub with all mushrooms highlighted while Photo 2 shows a graphic indicating picking sequence for those mushrooms — **Figure 2.** Example of picking sequence and bending direction within a mushroom tub. **(A)** mushrooms detected using a deep learning model, **(B)** determined picking sequence based on clustering and proximity to a defined origin point. Arrows indicate the optimal bending directions for detaching each targeted mushroom. Credit: Sadjad Mahnan

Mushroom Picking Mechanisms

Button mushrooms have delicate skin and are highly susceptible to bruising and mechanical damage, which can reduce their shelf life and market appeal. To address these challenges in robotic harvesting, two primary end-effector types have been explored: soft vacuum cups and finger grippers.

Vacuum Cups

Vacuum cups are widely used in robotic harvesting systems. However, lifting mushrooms vertically without incorporating bending and twisting motions requires greater suction force, which can lead to visible bruises or surface damage due to high vacuum pressure. Alternatively, when mushrooms are picked using bending, twisting, and lifting motions, the risk of injury remains if vacuum pressure is not properly adjusted or the cup material lacks sufficient softness (Figure 3). There are still challenges for vacuum cup-based mechanisms, while with carefully selected materials and cup design, robotic mushroom harvesting has great potential.

Photo 1 shows vacuum end-effector picking a mushroom. Photo 2 shows mushroom caps exhibiting bruising from excessive vacuum pressure. — **Figure 3.** Developed vacuum end-effectors designed for harvesting mushrooms through **(A)** bending motion. Credit: Mingsen Huang. **(B)** Bruises on mushroom caps caused by excessive vacuum pressure and insufficient softness of the vacuum cup during picking. Credit: Mingsen Huang.

Finger Gripper

To overcome the limitations of vacuum-based systems, researchers have explored the use of finger grippers for robotic mushroom harvesting systems. In this approach, soft, narrow fingers circle the mushroom cap from the side to achieve a secure grip. This design allows the mushroom to be detached using bending and lifting motions, or even solely by lifting without the risk of the cap detaching. However, the lateral gripping force can cause damage to the mushroom cap. Additionally, the mushroom's dense and tight growing environment makes it challenging to access the cap edges, limiting the applicability of this method.

Integrated Robotic Mushroom Harvesting Systems

A few integrated robotic prototypes have been developed to autonomously harvest mushrooms by companies such as 4AG Robotics (Vancouver, Canada). These systems typically include components for mushroom detection, selective picking through bending and twisting, followed by lifting motions, stem trimming, and placement into a collection container. To minimize the damage during harvesting, various designs of soft vacuum cups have been widely adopted in these systems.

Many other integrated robotic prototypes manufactured by companies such as Dorna Robotics (Upland, CA, USA) and Mycionics Inc. (Putnam, Canada) sought to develop robotic harvesting systems that mimic human fingers using two- and three-finger soft grippers for mushroom harvesting. These grippers, shaped to resemble mushroom caps, are capable of bending, twisting, and lifting mushrooms. These grippers are typically installed on a movable track on a mushroom shelf. The harvesting process is typically followed by stem trimming and packaging.

Additionally, some soft gripper designs integrate vacuum cups with finger grippers to leverage the advantages of both mechanisms.

Most existing robotic harvesting systems are designed specifically for Dutch aluminum shelves. However, these shelves are not affordable for many growers. Traditionally, many growers use wooden shelving systems, and since mushrooms do not require natural light, these shelves are often stacked closely together. As a result, there are structural constraints related to shelf height and aisle spacing in typical mushroom farms.

To accommodate these conditions, a mobile harvesting system will be a more practical solution. Such a system should be capable of navigating through narrow aisles, extending a manipulator into densely packed shelves, harvesting mushrooms from multiple levels, trimming their stems, and putting them into designated boxes.

Conclusion

Robotic mushroom harvesters are becoming more advanced and innovative through the improvement of mature mushroom detection, the adoption of the right picking motions, and the use of soft end-effectors. These systems can minimize the mushroom industry's dependency on labor. In the future, these systems will assist in the sustainable production of mushrooms by preserving the high quality of mushrooms and reducing the costs for growers.

Acknowledgements

This work was supported by the USDA NIFA Specialty Crop Research Initiative (SCRI) grant number 2021-51181-35859, and the Northeast Sustainable Agriculture Research and Education program under sub-award number GNE24-338. We would also like to convey our special thanks to Giorgio Mushroom Co. for their funding support through Penn State Mushroom Research Competitive Grant Program.