The images were gathered utilizing a simple time lapse camera to capture hours of images into brief video documents. Video is recorded at up to 1080p resolution. The cameras were introduced in the cranberry bogs all through the growing season catching pictures 24 x 7. The pictures are taken from various pieces of various swamps to catch the variability of the bogs. The pictures were taken in various circumstances like lighting conditions, points and directions. An IMU consists of six free sensors displayed on three symmetrical axes. On every one of the three axes is coupled an accelerometer and a gyroscope.

The main challenges in collecting the images data are: i) Traveling to field up and down for 10 hours two weeks once, ii) Weather condition was too cold in order to get into the cranberry marshes to collect the image from the mounted camera. iii) Battery life of the camera should be taken care of regularly for continuous image capturing. iv) Camera and electronics are prone to animals in the open field. v) Data management since there are too many images and it has to be in a sequential order.

The images are first manually classified in to different growth stages for training and testing the models. We considered totally 9 classes as different growth stages. This has been decided after talking to cranberry growers and understanding their needs for specific growth stages.

A big thanks to Dr. Mura lab, UW-Madison for allowing us to use the sample data for this project.