Abstract: In order to characterize the engineering rock mass authentically and completely, a new image recognition method is proposed relying on borehole camera technology. This method is based on Ant Colony Algorithm, and applied to the integrity evaluation of engineering rock mass. Firstly, in this image recognition method, the gray level is conversed and the target details is enhanced for borehole image, which facilitates the further identification and analysis of the rock mass characteristics. Secondly, the average joint width of the image is measured after the image segmentation based on Ant Colony Algorithm and morphological processing. And then the joints are divided into closed joint and open joint according to the joint aperture degree. Finally, the RQD of rock mass is calculated based on the identified rock mass characteristics. Take the drilling hole on the right bank of GX hydropower station on Lancang River as an example. The borehole digital imaging technology and RQD calculation were analyzed by this new image recognition method. Through the comparison with the core RQD value and wave velocity test results, this new image recognition method presents the following advantages of (ⅰ) fully considering the influence of joint aperture degree on physical and mechanical properties of rock mass; (ⅱ) reflecting the real topological state of engineering rock mass; (ⅲ) rationally evaluating the integrity of rock mass in areas with high ground stress, and (ⅳ) improving the scientificity and accuracy of RQD values. This work provides an efficient and reliable method for the integrity evaluation of rock mass.