RGB-Image-Based Forklift Pallet Recognition and Localization

Accurate identification and positioning of forklift pallets can effectively improve the automation level of warehousing logistics. Therefore, this paper proposes an algorithm based on SSD network and Canny edge detection operator to address the influence of image background and lighting changes on forklift pallet recognition. The pallet is simplified into an “E” model based on its structural features. The SSD network is utilized to detect and locate the “E” cross-section of the pallet, eliminating the interference of the main image background. The image is transformed into the HSV color space, and background interference is further eliminated by setting color thresholds. Gamma correction is applied to the V channel of the image in the HSV color space to eliminate lighting interference in pallet recognition. The Canny edge detection operator is used to detect the edges of the pallet and achieve segmentation of the “E” cross-section of the pallet. Pallet positioning is accomplished based on the segmentation results. Experimental results demonstrate that the proposed method can achieve pallet recognition and effective positioning in both bright and dark lighting environments.