Fishpond Visual Condition Dataset v2.0

Description

This dataset is a part of a fundamental research to produce an IoT monitoring device for fishpond. The hypothesis of this research is that the health of a fishpond can be inferred from the visual data. To build the dataset, several conditions data is gathered. The temperature, pH level, and total dissolved solid (TDS) were collected in several location at different time. At each time and location, an aerial photo of the pond was also collected using drone at several height. The images is presented in two condition: the raw original images of the ponds, and the cropped image on each data point. The conditions data is collected by using appropriate digital sensor for each parameter. The dataset consists of 975 data rows. Each row represent the condition and visual image (in 100 x 100 pixels images) of a fishpond at certain time and location. To use the dataset, please access the pond_dataset.csv file. The file contains the tabular data of 13 ponds (each pond represent different location and different collection time). For each row, the visual image file name is presented. To access the image file, please search in the images folder and find the corresponding image file according to the name listed in the csv file. The dataset can be used to study the correlation of each parameter. For example, the research originally study the correlation about the visual data with the conditions data. To do this, the image need to be preprocessed. The image data can be converted into a histogram data, or any other visual preprocessing method and result.

Steps to reproduce

This dataset is constructed using these instruments: 1. A digital thermometer for water usage/aquarium. 2. A digital TDS meter 3. A digital pH meter 4. Drone with camera To collect data on a fishpond, these steps were taken: 1. Choose a point in the fishpond. 2. Measure each condition parameter using measurement instrument. 3. Record the photo of fishpond spot using drone at 5, 10, 15, and 20 meter height. The photo need to be modified. The original size of the photo is usually in large resolutions. Modify the photo by cropping the photo at collection point so the size of the image will be 100 x 100 pixels.

Institutions

Categories

Funding

Licence