For precisely measuring the diameters of ruby balls in a non-contact manner， a compound secondary edge detection method is proposed. First， the image of the ruby ball is preprocessed， and the adaptive threshold Canny edge detection algorithm is used to detect the edge of the ruby ball image. Second， an image fusion algorithm based on the pixel weighted average is used to fuse the binary image of the ruby ball and the edge detection image. Then， the edge of the fused image is extracted， the cubic spline interpolation method is used to interpolate the edge image， the subpixel coordinates of the image edge are obtained via curve fitting， and circle fitting is performed according to the coordinates. Thereafter， the detection results are obtained through calibration. Compound secondary edge detection is performed for precisely measuring the diameter of a ruby ball. The experimental results indicate that the measurement accuracy for a 6-mm ruby ball diameter can reach 2 μm， and the positioning accuracy is less than 0.1 pixels， thus satisfying the measurement requirements of enterprises and providing better technical support for industrial automatic detection.