A Crop Canopy Localization Method Based on Ultrasonic Ranging and Iterative Self-Organizing Data Analysis Technique Algorithm

Abstract

To protect crops from diseases and increase yields, chemical agents are applied by boom sprayers. To achieve the optimal effect, the boom and the crop canopy should be kept at an appropriate distance. So, it is crucial to be able to distinguish the crop canopy from other plant leaves. Based on ultrasonic ranging, this paper adopts the fuzzy iterative self-organizing data analysis technique algorithm to identify the canopy location. According to the structural characteristics of the crop canopy, based on fuzzy clustering, the algorithm can dynamically adjust the number and center of clusters so as to get the optimal results. Therefore, the distances from the sensor to the canopy or the ground can be accurately acquired, and the influence of lower leaves on the measurement results can be alleviated. Potted corn plants from the 3-leaf stage to the 6-leaf stage were tested on an experiment bench. The results showed that the calculated distances from the sensor to the canopy using this method had good correlation with the manually measured distances. The maximum error of calculated values appeared at the 3-leaf stage. With the growth of plants, the error of calculated values decreased. The increased sensor moving speeds led to increased error due to the reduced data points. From the 3-leaf stage to the 5-leaf stage, the distances from the sensor to the ground can also be obtained at the same time. The method proposed in this paper provides a practical resolution to localize the canopy for adjusting the height of sprayer boom.

Keywords

• canopy localization
• fuzzy clustering
• self-organizing algorithm
• ultrasonic sensor
• sprayer boom height control