5. Results and discussions

In order to validate the effectiveness of the developed vision-based tactile detection method, an experiment is conducted to recognize a variety of shapes by using the proposed detection

Table 3. Detection results of variety of shapes.

algorithm. This experiment is conducted to prove that the proposed detection algorithm can compare different shapes such as circle, bar, eclipse, square, triangle, and diamond. These shapes have been captured through web camera and downloaded from Internet. A metric calculation is used to detect the shapes. The metric in the coding works by calculating any connected component's area and perimeter in a binary image after preprocessing, and then computing it using Eq. (1). After the metric has worked on the connected components, it will give a certain range of values for different shapes detected.

Table 3 shows the detection results of a variety of shapes by using the proposed tactile detection algorithm. From Table 3 results, the range of the metric value for each shape is confirmed by using the vision-based tactile detection algorithm. These metric values will be the benchmark value for each shape in order to differentiate the image shape. However, there are some shapes, which are having similar metric values such as circle, eclipse, and square. These analysis results will be used to improvise the current detection algorithm, making the system better and more robust to different types of detection environment of the tactile paving.

The detection of these various shapes will be used in order to recognize and differentiate the different shapes of items such as leaf, construction pavement, box, various shapes of papers/ garbage, etc. The detection of these various shapes will be used in order to recognize and differentiate the different shapes of items such as leaf, construction pavement, box, various shapes of papers/ garbage, etc. which usually covered the tactile pavement in real environment. A higher accuracy of detection system is needed to give higher reliability and give confidence to the visually impaired person when using the tactile pavement detection system to travel safely. Therefore, the result of the proposed detection system is very important for the blind navigation system, which is proposed in Figure 1.
