Development of 3 dimensional point cloud analysis technique to determine maturity of oil palm fresh fruit bunch based on frond position in oil palm phyllotaxis

A major challenge in oil palm industry is the shortage of labour mainly in harvesting operation. Harvesting operation accounts for 85% of total labour cost for oil palm sector showing that it is the crucial part of the operation in oil palm industry. Oil palm FFB need to be harvested at the right stage of ripeness to ensure optimum oil quality and quantity. Harvesters normally uses their experience in searching for loose fruits on the ground and observing the oil palm FFB’s color as an indication of matured bunch. In the early stage of this research, a new method to determine maturity stages of oil palm FFB based on its position in oil palm phyllotaxis frond spiral is verified. A formula to determine estimated harvesting date oil palm FFB called “Maturity Model” is derived based on the information of rate of frond production per month (RFP), duration of oil palm FFB development from anthesis (DOFFB) and position of oil palm FFB in frond spiral (FFBPFS). Oil palm FFB samples position in frond spiral is identified for harvesting date estimation and the FFB sample is monitored up to the harvesting day. In-field oil palm FFB sample date of harvest by farmers was compared with the estimated harvesting date calculated based on its position in frond spiral for validation. Regression analysis between Maturity Model estimated harvesting week and actual in-field harvesting week gives the coefficient of determination R2 = 0.90 and root mean square error (RMSE) of 1.58 weeks. From the result, the new method of determining maturity stages of oil palm FFB based on its position in frond spiral showed a positive outcome. Automatic approach to determine maturity based on frond position in Fibonacci phylloyaxis was introduced by using a 3D point cloud approach. The terrestrial LiDAR laser scanner (FARO Focus 3D 120) was used to acquire 3D point cloud data in order form a surface representation of oil palm tree. A Point Cloud Analyzation (PCLA) model was developed by using MATLAB application to analyze the characteristic of the oil palm tree. The PCLA model is able to identify the trunk diameter, spiral arrangement, total number of frond, label frond number based on phyllotaxis and determine the coordinate of each frond in world coordinate. Labelling frond number based on phyllotaxis is done based on tree height from the ground and then rectified using the Fibonacci phyllotaxis angle. PCLA model gives 76% accuracy for identification of frond spiral arrangement and coefficient of determination (R2) equal to 0.78 for identification total number of frond. A method to determine oil palm FFB was developed based on the height distribution of the fronds. Non-linear frond height distribution of the frond is a sign for the presence of the oil palm FFB. This is a potential new method of detecting the presence of oil palm FFB as well as the maturity stages. The outcome of this research will be helpful as a starter for a fully automated process to determine oil palm FFB maturity stages.

Preview Text FK 2018 118 IR.pdf Download (1MB) | Preview

Actions (login required)

View Item