Description

In some of the existing method is implemented the object recognition technique. But the limitations of the existing method are, the object is recognized only by the binary object. So to overcome these demerits, the LBP is implemented to get the pixel value of edge texture hence the object is recognized by the order of pixel value. In this proposed method, the LBP is implemented to recognize the object. Using INRAN data set. The input image is undergo for the preprocessing and then the image is then processed to separate the three channels known as the Red channel, Green channel and the Blue channel. The LBP technique is mainly focused to get the edge texture value of the image. Edge texture is nothing but getting the boundary value of the required object. At first the filter is used to de-noise the image using the medfilt2 filter. Then the canny filter is used to detect the boundary of the object, the weight of the pixel is noticed to the outermost boundary of the image. Each pixel is taken according to the coordinates of the x and y. when noticed the pixel value, the each pixel value is stored as an feature. The feature of the each pixel value is saved as a feature extraction. On the other hand, the train feature and the test feature is saved to match the trained image to the test image. The train feature and the test feature is match/classify using the classifier. The resultant data is further processed to classify the exact object.