Visual search is usually more efficient if the searched target was a salient item or on a salient location. However, our laboratory found a counterintuitive phenomenon that a target is actually more difficult to find if it was placed on a collinear salient structure in visual search, which is called the collinear masking effect. The collinear masking effect shows that a local target was more difficult to discriminate if the target was overlapping with the collinear structure, comparing to the target in the background. The collinear salient structure is the “super salient” item created according to the V1 model of salience by combining continuous perceptual grouping and feature contrast. Our studies showed that such a super salient item actually mask a local target not via perceptual salience but via collinear grouping. The ERP study revealed that the overlapping target reduced N2pc signals compare to the non-overlapping targets, suggesting the masking effect is perceptual. Also, the fMRI data showed that the critical region for the masking effect was on primary and associated visual cortex. Our work thus reveals the power of perceptual grouping that could lead to not only attentional capture but also camouflage.