Experimental Analysis of Line Detection Based on Image Segmentation

With the rapid development of computer technology, digital image analysis is increasingly becoming an important application in many fields such as face detection and, therefore, can be widely applied to various industries. However, while this technology aims to make people's lives more convenient, it still contains few issues about the precision of object detection. We proposed an experimental analysis based on image segmentation to tackle this issue, which can be used to explore the correlation between the precision of object detection and the adjusted threshold values. We use a multi-paradigm programming language MATLAB to extract the features of images. The process involves image denoising, sampling and quantization, segmentation, and compression. The experimental results demonstrated that the more sampling points surrounded around the shape of the objects represent, the better the line detection. Furthermore, we note that as the number of sample points decreases, the boundary of each pixel becomes clearer, which negatively affects the spatial resolution of the line detection image.