Blossom Detection Using Python and OpenCV
Additional Funding Sources
The project described was supported by a Specialty Crop Block Grant by the Idaho State Department of Agriculture.
Abstract
Having the ability to predict crop yield early into the season is crucial to farmers. One of the tools that can assist farmers in getting accurate early crop yield predictions is machine vision. Machine vision is adding a visual perception capability to a system a device to acquire data. The goal of this project is to develop a machine vision system to predict fruit yield by counting blossoms. The machine vision system used a color camera. The image processing algorithm was implemented using Python and the OpenCV library. Python is an open source language and is cost-effective option for farmers. The reason why the program was converted to Python was because of their open source language. The image processing used the color filters that were trained in last year’s data and was ran through this years images. The correlation between actual fruit count to blossom count were low as well as the correlation between expected fruit count to actual fruit count. These results show that there needs to be a modification to the color filter that was used as well as making sure the images are taken generally the same time of the year.
Blossom Detection Using Python and OpenCV
Having the ability to predict crop yield early into the season is crucial to farmers. One of the tools that can assist farmers in getting accurate early crop yield predictions is machine vision. Machine vision is adding a visual perception capability to a system a device to acquire data. The goal of this project is to develop a machine vision system to predict fruit yield by counting blossoms. The machine vision system used a color camera. The image processing algorithm was implemented using Python and the OpenCV library. Python is an open source language and is cost-effective option for farmers. The reason why the program was converted to Python was because of their open source language. The image processing used the color filters that were trained in last year’s data and was ran through this years images. The correlation between actual fruit count to blossom count were low as well as the correlation between expected fruit count to actual fruit count. These results show that there needs to be a modification to the color filter that was used as well as making sure the images are taken generally the same time of the year.
Comments
T23