Implementación de un prototipo de sistema para supervisión de lugares exclusivos para personas con discapacidad en parqueaderos públicos usando visión artificial.

Abstract

The objective of the work was to implement a prototype system for supervision of exclusive parking zones for people with disabilities in public parking lots using artificial vision. This system consists of a sensing stage, implemented by two pairs of E-R photoelectric sensors and linear displacement control for the camera, through a Nema 17 motor and an A4988 driver. This part performs the detection of the vehicle in each parking lot using sensors and consequently the positioning of the camera. Then the artificial vision stage, divided into two branches, the detection stage of the characteristic logo of people with disabilities and the detection of the vehicles license plates. Once the vehicle has been detected, shots are taken through the camera and images processing is carried out to recognize the logo and vehicle license plate, as required. The last stage consists of a warning alarm using a 6-tone standard siren, activating in the event that the characteristic logo is not recognized by the visual system. The solution for the problem of social assistance for people with disabilities was achieved, this through the tests carried out for each of the mentioned stages. For vehicle sensing, 94% and 96% reliability were achieved for sensor 1 and 2 respectively. For the evaluation of the visual system, 81.25% of effectiveness was achieved in the detection of the logo and 79.17% for the detection of plates, percentages being admitted, considering that there is an uncontrolled environment and lighting. The use of inductive loop detection, application of neural networks is recommended to improve system learning, sending information from plates to servers of the National Transit Agency (ANT) and alarm of custom voice commands.