Strawberry growers need to monitor pests to determine the options for pest management to reduce damage to yield and quality.  However, manually counting strawberry pests using a hand lens is time-consuming and biased by the observer. Therefore, an automated rapid pest scouting method in the strawberry field can save time and improve counting consistency. This study utilized six cameras to take images of the strawberry leaf. Due to the relatively small size of the strawberry pest, six cameras were used to take magnified images of the leaf; each camera can only cover a small portion of the leaf. The imaging box system is equipped with artificial illumination, which can reduce the negative effects of natural outdoor lighting. Strawberry leaf samples from ‘Sensation’ and ‘Brilliance’ cultivars were collected in the field and put inside the imaging box. Then, each of the six cameras collected ten images of single leaflets. After the image acquisition, the data was transferred to a desktop computer and a deep learning model was trained to detect the pests on the strawberry leaf. Currently, this study only collected predatory mites (Neoseiulus californicus and Phytoseiulus persimilis) images using the imaging system. Preliminary results had a pest detection accuracy (mean average precision) of 0.85. Two-spotted spider mite (TSSM) is the major pest of strawberries, and this study will collect many images of TSSM. The model detection performance is expected to be further improved, and TSSM will be detected after more images are collected and analyzed. This study demonstrated that the deep learning method could be successfully applied to the detection of very small arthropod pests with high detection accuracy. The deep learning model will be integrated with the imaging system to speed up strawberry pest detection in the field. The imaging system and deep learning method only need to scan the entire leaf once and can count the number of different pests simultaneously. This system will be much faster than manual counting.