How can I remove a picture's watermark using Matlab's image processing toolbox?

What are some effective object detection methods using image processing in MATLAB (Simulink)?

• I am currently working on a project of design ing a system which counts the moving large steel plates. Can anyone suggest effective methods of object detection using matlab image processing with simulink?

• Answer:

  It depends on the a priori characteristics of the object of interest against the rest of the environment. Normally, a combination of shape and color can be used, but it can be combined with other criteria such as: shape - rectangles can be detected by finding (approximately) vertical lines, (approximately) horizontal lines, and diagonal lines out of the edges in the image color - if the steel plates are silver, you can compute for the grayness of the object (e.g., if the RGB values are near, or if the Saturation is low) area (if the distance between the object and the camera is fixed, and if the size of the object is known) texture (This can be used to detect the small lines on the steel plates) compactness (The level of closeness to a perfect circle) number of holes (This should be zero for steel plates) If this is a real-time video application, I recommend C with Open Source Computer Vision Library, instead of MATLAB.