MYCOLab Ai Logo - gold and army green

Research on an Improved Segmentation Recognition Algorithm of Overlapping Agaricus bisporus

Summary

Scientists developed a new computer vision system that can automatically identify and locate overlapping button mushrooms (Agaricus bisporus) in factory farms. The system uses image processing techniques to overcome challenges like uneven lighting and crowded mushrooms. It successfully identified mushrooms with over 96% accuracy, which could help automate the mushroom harvesting process and reduce labor costs for farmers.

Background

Automated picking of Agaricus bisporus in factory environments faces challenges due to complex overlapping mushrooms and uneven illumination. Manual harvesting is labor-intensive, inefficient, and costly. Machine vision technology is essential for developing automated picking robots.

Objective

To develop an improved segmentation recognition algorithm that can accurately identify and segment overlapping Agaricus bisporus mushrooms in factory environments. The method aims to overcome illumination variance and complex adhesion between mushrooms.

Results

The method achieved a recognition rate of 98.81% with an average coordinate deviation rate of only 1.59%. The algorithm outperformed watershed and Hough circle transform methods, with recognition success rate of 97.25% and overall recognition success rate of 96.09%.

Conclusion

The proposed segmentation algorithm effectively segments overlapping Agaricus bisporus with recognition rates exceeding 96%, providing accurate picking information for automated harvesting robots. The method is adaptive to complex planting environments and uneven illumination, though computational efficiency could be improved.
  • Published in:Sensors (Basel),
  • Study Type:Algorithm Development and Experimental Validation,
  • Source: 10.3390/s22103946
Scroll to Top