In the process of using machine vision technology to detect the edges of circular holes and rectangular slots on furniture panel, the uneven stripes on the panel and the uneven distribution of details in the holes and slots will produce a lot of graphic noise, which will interfere with the edge information of holes and slots and affect the accuracy and stability of the detection results. This article proposes a method based on HSV color space segmentation, which can effectively eliminate unevenly distributed texture and detail noise and deepen edge information, and improve the reliability and accuracy of machine vision edge information. Firstly, the RGB raw image is converted to HSV color space image, and obtain the grayscale image with Hue, Saturation and Value channels. Then the hue channel grayscale image with obvious edge boundary between the hole, slot and the panel is selected, and the appropriate Gaussian kernel is selected for the image to perform Gaussian filtering. Finally, the edge of the filtered image is extracted by adjusting the appropriate upper and lower thresholds and using the Canny operator. The experimental results show that this method can effectively remove the texture of furniture panel and the noise of uneven details in the holes and slots, but also can almost retain the edge of the holes and the whole edge of the panel.