New Britain Palm Oil is making strides in the area of ‘precision agriculture’ in Papua New Guinea. Its experience provides pointers for other agribusinesses looking to gain a better control of their costs.

A view from above: a New Britain Palm Oil plantation. Credit: NBPOL

Papua New Guinea’s challenging terrain mean many farmers are flying blind when it comes to accurate data about their crops but new drone technologies can offer great advantages.

New Britain Palm Oil is PNG’s largest agricultural operator. At the end of 2018, the company had 11.36 million oil palms in PNG and the Solomon Islands.

Will Unsworth, Group TSD (Technology and Sustainable Development) Manager at New Britain Palm Oil Limited (NBPOL), says the company is aiming, with its aerial surveys of its operations, to achieve ‘precision agriculture’. This new method uses aerial technology to fully understand the business’s inputs, such as how much fertiliser it uses.

In 2018, he says, the company used 50,100 tonnes of fertiliser, costing US$20 million (K68 million).

‘Excess input is a waste; inadequate input is a loss. Precision avoids waste while maximising outputs,’ he notes.

‘After plenty of research and development, the company now has more accurate data than ever about how many oil palms it has and their distribution.’

Story continues after advertisment...

‘If fertiliser is over ordered, there are storage costs, more double-handling and losses from spoilage and theft,’ he says. ‘If the fertiliser is under-ordered, palms are under-fed, their yield falls, and there is a need to make additional orders, which can be costly.’

The company decided to better plan its use of fertiliser using the drone technology.

Aerial surveys

NBPOL’s Will Unsworth. Credit: Rocky Roe

Remote controlled drones are being used to generate aerial imagery of NBPOL’s plantations and to count in-field oil palms. From that, it is possible to generate an accurate ‘palm stand census’ to ensure that the correct fertiliser volumes are ordered.

NBPOL went through a series of steps:

• In 2016, it selected which drones to use
• In 2017, operators learned how to fly them
• In 2018, the aim was to learn how to process the data
• In 2019, the challenge was to learn how to deliver the data

Drones were selected based on cost, support and training, processing capabilities and the perceived quality of product.

Choices

Unsworth says there were differing opinions within the company about which direction to take with drones themselves, hardware and flight-planning software.

‘Fortunately, there were two groups which went with different options,’ he says, indicating that this divergence turned into a positive by identifying the strengths and weaknesses in both approaches. ‘Most of our expectations were wrong.’

One group chose a proprietary single software package that had multiple internal modules for each stage of the process. The other group preferred open source software.

Both software packages ‘suffered from continuous updates, in areas with limited internet access and this cost a lot of time,’ says Unsworth. ‘We used one system tied to the drone manufacturer, and one off-the-shelf graphics program for stitching together images.

‘They required immense amounts of checking and double-checking. There are false positives, false negatives, “whoopsies” and do-overs.’

Results

After plenty of research and development, the company now has more accurate data than ever about how many oil palms it has and their distribution. This data is shared using file sharing and converted into Google Earth for ease of access.

The project remains a work-in-progress. Unsworth says next step is to make further use of the data for surface modelling to improve water management and track the nutrient status of the oil palms from year-to-year.