Abstract: Image processing means is disclosed for calculating the position and orientation of an object with a known shape using a video signal from a TV-camera steadily aimed at the object. The means have a window memory into which are written several windows that are selected and shaped in advance, each of them covering a part of the picture received from the TV-camera. A window result memory has at least one memory cell per window and can be addressed by the window memory. An adder circuit is at its first input fed with a digital signal based on the in-signal of the TV-camera and at its second input with the contents of the memory cell of the window result memory corresponding to the window containing the pixel of current interest in the signal from the TV-camera. After this accumulating adding, the adder circuit returns the result to the same memory cell in the window result memory.