Georgia’s peanut crop brings in over $ 600 million a year, but figuring out the optimal harvest time can be tricky. Modern yield estimation tools are imperfect and their work results in huge yield losses every year. In a new study published in Springer Precision Agriculture, researchers at the University of Georgia talked about a new portable smartphone-based system for analyzing peanut crops.
To determine whether peanuts are ripe or not, farmers use a special grading system using a profile board and digging samples from their field to assess the state of the crop by the county manager. The day of harvest of the peanuts is set according to the color of the dug peanuts. Such a system is too dependent on the human factor and takes a long time.
The new system works as follows.
The farmer places the samples on the profile board by placing the peanut samples in separate slots. After that, the board must be inserted into the lower part of the portable photo station developed by the scientists. By placing the smartphone on top, the user takes a picture of the sample. The mobile application analyzes the color of peanuts and feeds the information to an online database. It also allows the user to customize their report for the district manager by entering information such as field geodata, peanut type, and overall crop status.
This faster assessment tool will allow farmers to constantly monitor the maturity of their crops and take into account factors such as soil type and weather, the researchers said.
“Unripe or overripe peanuts can have high levels of aflatoxin, a toxin produced by fungi found on some crops,” explains Rui Li, lead author of the study and a graduate of the University of Georgia. If the farmer can continuously monitor the maturity level of the peanuts, he can start the harvesting process at the right time and reduce the yield losses caused by aflatoxin.”
The researchers conducted 52 tests onsite to test the accuracy of the system. Compared to people assessing the harvest, the system showed much fewer errors than the human.
The system can be improved in two ways, scientists say. One is to further optimize the lighting conditions to ensure image quality, and the other is to improve the color detection function using a special learning algorithm.
Once enough data has been entered into the database for evaluation, machine learning algorithms will improve the accuracy of the system.
The development of this system will not only significantly save the time of the peanut farmers, but also improve the quality of the peanuts for the consumer.
Don Koehler, executive director of the Georgia Peanut Commission
“The farmers (who participated in the field trials) were strongly interested in the benefits of the smartphone app. They love the portability and low cost of the system, ”Lee concludes.