Line Segment Detection

Characteristics such as plant height and branch length are critical for crop production and mechanization of harvesting processes. To facilitate automatic measurement of these parameters in rapeseed plants, we developed a method that employs image processing to detect line segments within photographs of plants. Our approach incorporates the Line Segment Detector algorithm (Von Gioi et al., 2008) — a method recognized for its speed, precision, and leading-edge performance in line detection. To refine the output and remove extraneous and broken line segments, we implemented a series of filtering strategies that leverage the geometric properties of the detected lines. Experimental results reveal an impressively low error margin of 1.25% while utilizing modest computational resources, thereby aligning closely with the practical requirements of agricultural applications.