Yield prediction is an increasingly researched topic aimed at boosting productivity in agricultural settings. Leveraging recent advancements in AI, particularly in Image Processing and Computer Vision, our team focuses on innovative approaches to improve yield prediction in blueberry crops. Besides, we are interested in all the pipeline involved for predictions from devices for autonomously collecting data, field mapping and identification, to the data representation and further temporal analyses.
Blueberries present unique challenges due to their intricate plant morphology. The combination of small berry size, abundant quantity, high leaf density, and varied canopies results in numerous occlusions and difficulties in detection. Our team’s primary goal is to enhance prediction through the integration of cutting-edge computer vision techniques and on-site implementations to adapting solutions to the agricultural scenario.
Research Focus:
Our team concerned on improving the following key aspects:
Detecting Berries in the Bush: Our research involves developing robust algorithms for accurately detecting berries amidst complex foliage, thus enabling more efficient yield estimation.
Optimal Harvest Point Forecasting: By providing farmers with reliable predictions of the optimal harvest point, our solutions can significantly improve overall productivity.
Resource Allocation Support: The estimated berry quantities aid farmers in allocating human resources and other essential inputs more effectively during harvesting.
Phenotyping for Plant Breeders: Our work also extends to facilitating the decision-making process for plant breeders by enabling precise phenotyping of each cultivar.
Potential Applications:
The outcomes of our research have various practical applications, including:
- Research and Breeding: Assisting plant breeders in selecting superior cultivars and enhancing breeding programs.
- Informed decisions: Empowering farmers to make informed decisions and maximize yields.
- Forecasting: Facilitating efficient resource allocation, leading to cost reduction and increased sustainability.
Our team is committed to pushing the boundaries of “Field High-Throughput Phenomics” by embracing emerging technologies and fostering collaborations within the scientific community. If you’re interested in our research or would like to collaborate, please feel free to reach out to us.