Objectives
To optimize palm oil extraction yield by implementing an automated FFB (Fresh Fruit Bunch) grading system that accurately assesses fruit ripeness and quality, reducing human error and increasing operational efficiency.
Problem statement
Manual FFB grading processes are susceptible to human error, leading to inconsistent and potentially biased assessments. The practice of conducting FFB grading twice, at both the estate and mill levels, introduces further discrepancies and delays in the production process. These inefficiencies result in suboptimal palm oil extraction rates and disputes between stakeholders.
Solutions
Develop an automated FFB grading system utilising a light normalisation module and deep learning algorithms for ripeness estimation. The system accurately assesses FFB quality parameters such as colour, size, and shape, providing real-time data for decision-making. By eliminating human intervention and standardising the grading process, the system improves the precision and efficiency of FFB classification.
Results/Benefits
- Optimised Oil Extraction: Accurate FFB grading enables better matching of fruit ripeness to extraction processes, maximizing oil yield (8-12% improvement from baseline oil extraction rate or OER).
- Enhanced Efficiency: Automated grading streamlines the FFB handling process, reducing labour costs and processing time.
- Improved Accuracy: Eliminates human error in FFB assessment, leading to more precise and consistent grading.
- Data-Driven Decision Making: Provides valuable data on FFB quality and characteristics for process optimisation and quality control.
- Dispute Resolution: Reduces disputes between estate and mill due to standardised grading criteria.
Increased Profitability: By improving oil extraction efficiency and reducing costs, the system contributes to overall profitability.
