Case studies

Calculating brix & yield in sugarcane with RedEdge data


Share | 04/30/2021

HG Robotics team getting ready to fly a sugarcane field.

HG Robotics Co., Ltd. (HGR), in partnership with BSFR (Bio-Sensing and Field Robotics Laboratory), Khon Kaen University, and Global Crops Co., Ltd., is trying to improve sugarcane production in Thailand with the use of multispectral technology and without disrupting traditional practices.

Brix & yield

While for many crops yield is the primary indicator of a successful season, in sugarcane production crop yield is not enough, the Brix value of sugarcane juice is another critical parameter. Farmers are being paid based on the percentage of sugar and not the number of canes produced. More stalks with high Brix means more money for the producer.

“Farmers don’t want to combine their cane to others even when their plots are next to each other because the price of the sugarcane also depends on the sweetness,” Santawat Santiteerakul, Product Development Manager at HG Robotics.

HG Robotics solution aims to address these two issues. With more accurate data, farmers will be better positioned to negotiate with the sugar factories and be able to combine their cane and still get fair benefits. Knowing the average Brix and expected yield for their plots, farmers will be able to harvest both fields together and split the benefit proportionally based on the Brix and yield values of each plot.

The solution

HG Robotics team member capturing an image of the MicaSense series CRP before flying to ensure radiometric calibration.

Using the MicaSense series RedEdge and their own UAV, the HG Robotics Vespa Hex, HG Robotics has been flying over Khon Kaen University research fields in northeastern Thailand and a 160 hectares farm in the Khon Kaen province four times a year for the past four years, aiming to obtain high-quality data on both Brix and yield.

Having a cost-effective way to estimate both Brix and yield will also impact the bottom line by reducing the logistics cost since it would allow multiple plots to be harvested simultaneously.

In addition, the crop health maps generated from the RedEdge multispectral data are allowing for targeted applications of herbicides and fertilizer. This means that by having localized applications, farmers can reduce the amount of product used, leading to a reduced cost of inputs.

The analysis

HG Robotics has developed an automated model for processing the data collected, which uses both Agisoft Metashape and QGIS and takes about four minutes to process a hectare. 

This analysis has led the team to find a strong correlation between the Brix values and the NIR and Red Edge reflectance values, allowing them to calculate the Brix of the crops after each flight and compare the results with those obtained from previous flights. 

“Without RedEdge data, our models basically cannot calculate the Brix value at all. Apart from that, our models require high-resolution data,” Santawat Santiteerakul, Product Development Manager at HG Robotics.

The results

RGB images from the MicaSense series RedEdge showing a time comparison of the field from October to December. The one on the left shows the extent of the damage. HG Robotics using Hivegrid, their processing solution, was able to make yield and Brix predictions that led to earlier harvesting.

In 2019 Thailand was hit with the worst drought in the last 40 years. The lack of proper irrigation significantly affected the sugarcane fields resulting in a yield decrease of 20% in Q4 compared to Q3. HG Robotics’ Hivegrid was put to the test. The extent of damage from water stress was quantified through their drone mapping and processing workflow. The data suggested that an early harvest would be favorable to avoid further water stress damage to the crop. The early harvest helped salvage about 200 tonnes of sugarcane (~6% of the total production of the season) which translated into $6,900 saved.

To learn more about the applications of multispectral imagery in sugar cane production, watch HG Robotics & this webinar and the sensors page here.

Newsletter sign up

Sign up to receive updates directly to your inbox.