A system for detecting and tracking a curvilinear object in a three-dimensional space includes an image acquisition system including a video camera arranged to acquire a video image of the curvilinear object and output a corresponding video signal, the video image comprising a plurality n of image frames each at a respective time t.sub.i, where i=1, 2, . . . , n; and a data processing system adapted to communicate with the image acquisition system to receive the video signal. The data processing system is configured to determine a position, orientation and shape of the curvilinear object in the three-dimensional space at each time t.sub.i by forming a computational model of the curvilinear object at each time t.sub.i such that a projection of the computation model of the curvilinear object at each time ti onto a corresponding frame of the plurality of image frames of the video image matches a curvilinear image in the frame to a predetermined accuracy to thereby detect and track the curvilinear object from time t.sub.1 to time t.sub.n.

An edge extraction unit extracts an edge image from a photographed image obtained by photographing a product with a camera. A constant edge acquisition unit acquires, as a constant edge image, an edge image in a constant surface where a positional deviation does not occur with respect to a welding point set by a processing program, the acquired image belonging to the extracted edge image. A correction amount acquisition unit performs pattern-matching between a master constant edge image and a workpiece edge image, which are acquired by the constant edge acquisition unit, and acquires a deviation amount between both thereof as a correction amount with respect to the welding point. A processing program correction unit corrects the welding point by the correction amount, and generates a corrected processing program for welding the workpiece. A welding robot welds the workpiece based on the corrected processing program.

A robot control method includes a teaching step, first processing step, modifying step, second processing step, and third processing step. In the modifying step, a third teaching point is changed to a second modified point, a fourth teaching point to a third modified point, and a fifth teaching point to a fourth modified point, based on a difference between a second teaching point and a first modified point. A profile modifying control to change the position of a work tool is applied, using a sensor mounted on the processing advancing direction side of the work tool, in the first processing step and the third processing step. An attitude of the work tool is changed during the second processing step.

A system for detecting and tracking a curvilinear object in a three-dimensional space includes an image acquisition system including a video camera arranged to acquire a video image of the curvilinear object and output a corresponding video signal, the video image comprising a plurality n of image frames each at a respective time t.sub.i, where i=1, 2, . . . , n; and a data processing system adapted to communicate with the image acquisition system to receive the video signal. The data processing system is configured to determine a position, orientation and shape of the curvilinear object in the three-dimensional space at each time t.sub.i by forming a computational model of the curvilinear object at each time t.sub.i such that a projection of the computation model of the curvilinear object at each time ti onto a corresponding frame of the plurality of image frames of the video image matches a curvilinear image in the frame to a predetermined accuracy to thereby detect and track the curvilinear object from time t.sub.1 to time t.sub.n.