SE-COTR: A Novel Fruit Segmentation Model for Green Apples Application in Complex Orchard.
Zhifen Wang1, Zhonghua Zhang1, Yuqi Lu1
1School of Information Science and Engineering, Shandong Normal University, Jinan 250358, China.
Plant Phenomics (Washington, D.C.)
|June 2, 2023
Summary
A new deep learning method, SE-COTR, accurately segments green apples in complex orchards. This innovative approach improves fruit detection, even for small or occluded targets, enabling efficient agricultural applications.
Area of Science:
- Agricultural Engineering
- Computer Vision
- Artificial Intelligence
Background:
- Efficient green fruit detection and segmentation in natural orchards is challenging due to unstructured environments.
- Existing methods struggle with varying fruit sizes, occlusion, and similar background colors.
Purpose of the Study:
- To propose an innovative deep learning method, SE-COTR (segmentation based on coordinate transformer), for accurate and real-time green apple segmentation.
- To enhance feature focus and integrate multiscale features for improved segmentation performance in complex orchard conditions.
Main Methods:
- Utilized MobileNetV2 as a lightweight backbone.
- Developed a coordinate attention-based coordinate transformer module to enhance feature focus.
- Implemented a joint pyramid upsampling module for multiscale feature integration.
- Applied dynamic convolution for instance mask prediction.
Main Results:
- SE-COTR achieved a mean average precision of 61.6% for green apple segmentation in complex orchards with occlusion and varying scales.
- Segmentation accuracy for small target fruits reached 43.3%, outperforming advanced models.
- The method demonstrated low complexity and effectiveness in challenging environments.
Conclusions:
- SE-COTR effectively addresses low accuracy and model complexity issues in green fruit segmentation.
- The model can be deployed on portable devices for accurate and efficient agricultural tasks in complex orchards.
- The proposed method offers a significant advancement for intelligent agriculture and automated fruit harvesting.
Related Concept Videos
01:58Fruit Development, Structure, and Function
22.5K
Fruits form from a mature flower ovary. As seeds develop from the ovules contained within, the ovary wall undergoes a series of complex changes to form fruit. In some fruits, such as soybeans, the ovary wall dries; in other fruits, such as grapes, it remains fleshy. In some cases, organs other than the ovary contribute to fruit formation; such fruits are called accessory fruits.
22.5K
02:16Light Acquisition
8.5K
In order to produce glucose, plants need to capture sufficient light energy. Many modern plants have evolved leaves specialized for light acquisition. Leaves can be only millimeters in width or tens of meters wide, depending on the environment. Due to competition for sunlight, evolution has driven the evolution of increasingly larger leaves and taller plants, to avoid shading by their neighbors with contaminant elaboration of root architecture and mechanisms to transport water and nutrients.
8.5K