| The research of this paper is focused on the application of noise image segementation to the multi-function road surface status inspection vehicle and the intelligent vehicle. Based on the engineering background, the methods of transportation imges denoising, the road crack detection, lane detection and motion detection is studied. The main contents and contributions of this dissertation are as follows.1. Image denoise technology.A new impulse noise detector based on an adaptive neuro-fuzzy inference system (ANFIS) is presented. The proposed operator is a hybrid filter obtained by appropriately combining a median filtering, a wiener filtering and the ANFIS. The noise is exactly estimated through the proposed operator. The internal parameters of the ANFIS are adaptively optimized by training. The training is easily accomplished by using simple artificial images that can be generated in a computer. The distinctive feature of the proposed operator is that it offers well line, edge, detail, and texture preservation performance while, at the same time, effectively removing noise from the input image. Simulation experiments show that the proposed operator may be used for efficient restoration of digital images corrupted by impulse noise without distorting the useful information in the image.2. Road crack detection technology.We proposed a pavement distress detection algorithm based on tiles in wavelet domain. After the preprocesses of de-noise, enhancing, etc., pavement image was decomposed by Haar wavelet and the approximation coefficient in the highest layer was segmented by adaptive threshold to get the initial pavement distress regions. Then in the other ordinal layers, from the higher layer to the lower layer, only the pavement distress regions were segmented tiles to gain the exact detection of pavement distress. Simulation show the algorithm proposed in this paper could detect pavement distress effectively for different types and could overcome the effect of the noisy.3. lane detection technology.In the lane detection, we propose a method based on visual moden. The thresholds of segementation are calculated according to Weber's law and bionics principle. The lane detection fuses the way of lane edges detection and the method of color-based segmentation. The method reduces the shadow and noise points influnce on lane detection and is convencient in the applications in the vision navigation of the autonomous highway vehicle. The simulation results reveals that our method gets a better result in different road conditions.4. edge detection technology.Neuro-fuzzy (NF) systems are very suitable tools to deal with uncertainty encountered in the process of extracting useful information from images. We present a novel adaptive neuro-fuzzy inference system (ANFIS) for edge detection in digital images. The internal parameters of the proposed ANFIS edge detector are optimized by training using very simple artificial images. The edges are directly determined by ANFIS network. The proposed ANFIS edge detector is tested on popular images having different image properties and also compared with popular edge detectors from the literature. Experimental results show that the proposed ANFIS edge detector exhibits much better performance than the competing operators and may efficiently be used for the detection of edges in digital images.5. background reconstruction technology.A new background subtraction algorithm based on two thresholds sequential clustering is proposed in this paper. First, pixel intensity in period of time is classified based on two thresholds sequential clustering. Second, combination process is run to classified classes. Finally, the backgrounds of scene are selected, so the background model can represent the scene well. The simulation results show that the proposed algorithm is robust to the thresholds, those near classes are avoided at all, and the effect of input order of data has been reduced greatly. And the background model can represent the scene well.
|
PDF Full Text Request