Prototipo de sistema de visión artificial en un vehículo aéreo no tripulado para la detección de órdenes gestuales

The main objective of this degree project is to facilitate the navigation of the operator of an unmanned aerial vessel through a camera and using an artificial vision system will be able to recognize patterns made with the back and arm of a person fully identified. Using quadripopter Ar dron 2.0, a control system based on body gestures was developed that allows handling translation movements using artificial vision. The titling work allows to identify the gestures made with the torso and arm of a person captured by the camera on board and transmitted to a fixed station, that is to say, a portable computer that allows to process the video flow making use of the Matlab software. The method based on the appearance of the image allows selecting the appropriate commands that were used for the creation of the database and the design of the control system. The work carried out consists of three events: take off where the central pixels of the image are accessed and wait for the RGB components to have a value of zero, navigation is used to adapt the correlation method to identify the patterns and landing where it extracts the coordinates of the centroid of the operator and if they leave the hysteresis zone it proceeds to perform said action. The results and feasibility of the developed system are checked by several users comparing the proposed project with the conventional system allowing observing and documenting the versatility of the project.

• Tác giả chính: Bustillos Catota, Ana Belen (author)
• Tác giả khác: Farinango Defaz, Daniel Patricio (author)
• Định dạng: bachelorThesis
• Ngôn ngữ: spa
• Được phát hành: 2018
• Những chủ đề: PROGRAMAS PARA COMPUTADOR; TELECOMUNICACIONES; CONTROL AUTOMÁTICO; VEHÍCULOS AÉREOS NO TRIPULADOS
• Truy cập trực tuyến: http://dspace.ups.edu.ec/handle/123456789/15225