Navigation of autonomous vehicles in natural environments based on image processing is certainly a complex problem due to the dynamic characteristics of aquatic surfaces, such as brightness and color saturation. This paper presents a new approach to identify turbid water surfaces based on their optical properties, aiming to allow automatic navigation of autonomous vehicles regarding inspection, mitigation and management of aquatic natural disasters. More speciﬁcally, computer vision techniques were employed in conjunction to artiﬁcial neural networks (ANNs), in order to build a classiﬁer designed to generate a navigation map that is interpreted by a state machine for decision making. To do so, a study on the use of different features based on color and texture of such turbid surfaces was conducted. In order to compress the extracted information, Principal Component Analysis (PCA) was performed and its results were used as inputs to ANN. The whole developed approach was embedded in an aquatic vehicle, and results and assessments were validated in real environments and different scenarios.
In order to evaluate the developed approach for autonomous navigation in a real environment, it was embedded in an aquatic vehicle. We build and develop our approach using the programming language C, with support of OpenCV library, OpenMP for multiprocessing programming, and Fast Artiﬁcial Neural Network Library (FANN), a free library that implements an ANN multilayer in language C . The hardware used was a Raspberry Pi board (RPI) model 2 and a Raspberry camera. Figure 11-(b) presents the prototype of the aquatic vehicle with the RPI board and camera connected.
Mateus Eugênio Colet, Adriana Braun, Isabel H. Manssour
PUCRS, Faculdade de Informática
Porto Alegre, RS, Brazil