Joanne the farmer sets out for an action packed day. She wants to collect data to help predict which apples from each cluster will last and which will fall from the tree. She will be collecting data regarding the growth of each apple, and the relative growth rate of each apple compared to its cluster. To do this, she will be using the Orchard Watch app. At each cluster, she will open the camera feature inside the app, and will slowly move the phone as the camera is pointed at the AprilTag. As she moves the phone, changing its orientation in regards to the AprilTag and the cluster, the App will take a picture when the phone has reached the desired location to make accurate measurements.
Once the picture is taken, the image will be sent to the Orchard Watch server for ML analysis. The ML model will compare the relative size of each fruitlet with the constant-width AprilTag. The model calculates the pixel width of the fruitlet in the image and compares it with the AprilTag. A ratio of the fruitlet and AprilTag is stored as the fruitlet's original width, along with the photo's timestamp for each uniquely identified fruitlet. Joanne’s day is off to a good start, as she really enjoys not having to measure every apple herself.
Next week, she will return to the same tree and photograph the same cluster. The same model will calculate the width of each fruitlet and properly store it. Using the widths from last week, the app is able to tell Joanne how much each fruitlet has grown or shrunk. Over time, enough growth data will be collected to use Researcher Paul's Fruitlet Growth Model to Predict Thinner Responses on Apples. Then, the app will predict the thinner response based on fruitlet growth.
Image uploading is related to #29. Data analysis is related to #98.
Joanne the farmer sets out for an action packed day. She wants to collect data to help predict which apples from each cluster will last and which will fall from the tree. She will be collecting data regarding the growth of each apple, and the relative growth rate of each apple compared to its cluster. To do this, she will be using the Orchard Watch app. At each cluster, she will open the camera feature inside the app, and will slowly move the phone as the camera is pointed at the AprilTag. As she moves the phone, changing its orientation in regards to the AprilTag and the cluster, the App will take a picture when the phone has reached the desired location to make accurate measurements.
Once the picture is taken, the image will be sent to the Orchard Watch server for ML analysis. The ML model will compare the relative size of each fruitlet with the constant-width AprilTag. The model calculates the pixel width of the fruitlet in the image and compares it with the AprilTag. A ratio of the fruitlet and AprilTag is stored as the fruitlet's original width, along with the photo's timestamp for each uniquely identified fruitlet. Joanne’s day is off to a good start, as she really enjoys not having to measure every apple herself.
Next week, she will return to the same tree and photograph the same cluster. The same model will calculate the width of each fruitlet and properly store it. Using the widths from last week, the app is able to tell Joanne how much each fruitlet has grown or shrunk. Over time, enough growth data will be collected to use Researcher Paul's Fruitlet Growth Model to Predict Thinner Responses on Apples. Then, the app will predict the thinner response based on fruitlet growth.
Image uploading is related to #29. Data analysis is related to #98.