In this paper, the authors develop the weed detection system and the classification algorithm which is applied for autonomous weed controller robot. To achieve this objective by using the variety of machine vision algorithm the weeds and crops were classified between plants. Weed is a plants which is grow in between the row, crops was mixed with weeds. So classification techniques are required to classify the weeds from the crop. According to that a features extracted by the pattern matching analysis of the image. Once the weed position was identified and the robots spray the herbicide on the position of a weed.