Related Experiment Video
Updated: Jun 14, 2025
A Pipeline using Bilateral In Utero Electroporation to Interrogate Genetic Influences on Rodent Behavior
Published on: May 21, 2020
Screening of identification algorithm for rodent-induced bare patches based on the drone imagery
Bin Cai1,2, Rui Dong1,2, Rui Hua3
1College of Pratacultural Science, Gansu Agricultural Universit/Key Laboratory of Grassland Ecosystems of the Mini-stry of Education/Engineering and Technology Research Center for Alpine Rodent Pest Control, National Forestry and Grassland Administration, Lanzhou 730070, China.
Unmanned Aerial Vehicle (UAV) remote sensing accurately identifies rodent-damaged bald spots in alpine meadows. The neural network (NN) classification method proved highly effective for ecological hazard assessment.
Area of Science:
- Ecology
- Remote Sensing
- Geospatial Analysis
Background:
- Rodent-infested bald spots are key indicators of grassland degradation.
- Plateau pika activity creates these detrimental habitats, posing ecological risks.
- Unmanned Aerial Vehicle (UAV) remote sensing offers a novel approach for monitoring these areas.
Purpose of the Study:
- To evaluate supervised classification algorithms for identifying plateau pika habitats from UAV imagery.
- To determine the most accurate method for mapping rodent-damaged bald spots in alpine meadows.
Main Methods:
- Utilized UAV-visible light imagery for data acquisition.
- Applied five supervised classification algorithms: minimum distance (MinD), maximum likelihood (ML), support vector machine (SVM), Mahalanobis distance (MD), and neural network (NN).
- Assessed classification accuracy using confusion matrices and Kappa coefficients.
Main Results:
- Neural network (NN) and support vector machine (SVM) classification showed superior performance.
- NN achieved high mapping accuracy for grassland (98.1%) and bald spots (98.5%).
- The NN model demonstrated overall accuracy of 98.3% with a Kappa coefficient of 0.97, indicating excellent agreement and stability.
Conclusions:
- The neural network (NN) method is highly suitable for identifying rodent-damaged bald spots in alpine meadows using UAV remote sensing.
- This technology provides a reliable tool for ecological hazard assessment related to rodent infestations.
- Accurate mapping of these habitats is crucial for effective grassland management and conservation efforts.
