Abstract: To judge the mixing uniformity of powder materials during the preparation of sintered abrasive blocks, a detection method of powder uniformity based on the similarity of multiple powder imagesis proposed. Firstly, a deviceis designed to collect powder samples online, the image of which is collected by digital microscope. Then, the statistical features of gray distribution based on gray histogram are extracted as the feature parameters of the collected images. Principal component analysis (PCA) is used to reduce the dimension of image features and to construct image feature matrix. Finally the variable weight similarity of each powder image tothe sample density center in each set of images is calculated. The smallest one taken as the similarity with an appropriate thresholdis set to judge the powder whether to mix evenly. The experimental results show that the proposed method can accurately judge the mixing endpoint and determine the optimal mixing time. It is suitable for judging the mixing uniformity of powder materials in the sintered abrasive blocks. It is helpful to improve production efficiency and grinding block quality andincrease economic benefit.