Published: 06-07-2020| Version 6 | DOI: 10.17632/8ysxtyf8p2.6
Vanessa Ap. de Moraes Weber,
Edson Matsubara,
Urbano Gomes P. de Abreu,
Hemerson Pistori


Image Dataset with 51 classes the Pantaneiro Cattle. The images in the dataset were collected between 1:50pm and 4:48pm on January 25, 2019 at the Aquidauana Pantanal Cattle Conservation Center (NUBOPAN) in the State University of Mato Grosso do Sul/Aquidauana University Unit (UEMS/UUA). This is located in the municipality of Aquidauan/MS which is in the Pantanal region of Mato Grosso do Sul. The geographical coordinates were as follows: latitude 20º28' S, longitude 55º48' W and an altitude of 149 meters. In this study, 51 Pantanal cattle units of both sexes and various ages were used. This is equivalent to 10% of the existing population for this breed. Images were collected using the Digital Video Recorder (DVR) set MIDI MD-1004NS MD-DVR41. The image quality was AHD 720p and the recording capacity was 1Tb. Four cameras were positioned at the exit outlet to collect images of the back, profile, side, and face. Images of the cattle’s backs were collected using camera 1. This was positioned approximately 2.5 meters high with an inclination angle of 30º. The other cameras were positioned according to the sketch at the same height as the pickets. The videos collected were stored in the DVR, transferred to computers, analyzed, and divided into 212 smaller videos. Images of the cattle were then extracted from these videos. The dataset was formed using 27,849 images of Pantanal cattle and the number of images obtained varied according to filming to time of each cattle.