EcoCropsAID: Thailand’s Economic Crops Aerial Image Dataset

Thailand’s Economic Crops Aerial Image Dataset (EcoCropsAID)

We introduce the novel Economic Crops Aerial Image Dataset, named the EcoCropsAID dataset. This dataset was collected in Thailand and consists of aerial images of five key economic crops cultivated in different provinces and regions between 2014 and 2018. The dataset was compiled based on data from Agri-Map Online, a tool provided by the Ministry of Agriculture and Cooperatives and the National Electronics and Computer Technology Center (NECTEC). Agri-Map Online is an agricultural mapping system widely used by departments under the Ministry of Agriculture and Cooperatives for agricultural management. It ensures that the agricultural information is accurate and up-to-date. The Google Earth application was subsequently employed to capture aerial images of the selected economic crop areas. The dataset is complex, as the Google Earth program utilizes various remote imaging sensors to capture the aerial images.

The EcoCropsAID dataset includes five categories (rice, sugarcane, cassava, rubber, and longan), containing a total of 5,400 images, with each class comprising approximately 1,000 images. The aerial images were captured at a resolution of 600 × 600 pixels and stored in RGB color format.

The challenges associated with classifying the EcoCropsAID dataset include: 1) the presence of varying image resolutions and colors due to the different remote imaging sensors used, 2) the similarity of patterns among certain classes, such as longan and rubber, and 3) the differences in patterns within the same class, such as cassava and rice.