Detection of 3-D objects with a 2-D laser scanning sensor for greenhouse spray applications

Authors: YAN, TINGTING - Nanjing Agricultural University; Zhu, Heping; SUN, LI - Jiangsu University; WANG, XIAOCHAN - Nanjing Agricultural University; LING, PETER - The Ohio State University

Submitted to: Computers and Electronics in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/30/2018
Publication Date: 12/10/2018
Publication URL: https://handle.nal.usda.gov/10113/6101615
Citation: Yan, T., Zhu, H., Sun, L., Wang, X., Ling, P. 2018. Detection of 3-D objects with a 2-D laser scanning sensor for greenhouse spray applications. Computers and Electronics in Agriculture. 152:363-374. https://doi.org/10.1016/j.compag.2018.07.030.
DOI: https://doi.org/10.1016/j.compag.2018.07.030

Interpretive Summary: Greenhouse crops are produced in confined biophysical systems with artificially controlled inputs and outputs, which requires intensive use of pesticides, fungicides, fertilizers and water to achieve the optimal plant growth during their entire life cycle. Precision spray technology is needed to prevent greenhouse production from excessive waste of pesticide, nutrition and water. Variable-rate technology to automatically control spray outputs to target areas may be a potential solution for greenhouse applications, and sensors are the critical component of the precision sprayers. The goal of this research was to validate the accuracy and capability of an inexpensive indoor-use laser sensor in detecting complex-shaped plants before this sensor technology could be integrated into the variable-rate spraying systems for greenhouse applications. Validations included: 1) using the laser sensor to detect four different regular-shape objects with different color and surface finish and two different-sized artificial plants under different sensor travel speed, detection height and object position conditions; 2) calculating the object dimensions by a specially designed algorithm; 3) validating the measurement accuracy of the laser sensor with the root mean square error and coefficient of variation, as well as the pseudo-color images mapping the 3-D object surface; 4) calculating the theoretical detection errors to ratify the feasibility of implementing the laser sensor to the greenhouse spray application. These validations provided a baseline for the laser sensor application under indoor simulated greenhouse environments.

Technical Abstract: Precise measurement of plant dimensions can provide essential information for variable-rate spray application technologies to discharge the amounts of chemicals actually needed by the plants. The accuracy of a 270° radial laser scanning sensor combined with a specially-designed plant surface mapping algorithm was evaluated in detection of complex-shaped object surfaces and sizes. Data acquisition and three-dimensional (3-D) image construction were supported with the specially-designed algorithm. The objects in the test were toy balls with a pink smooth surface, light brown rectangular cardboard boxes, black and red texture surfaced basketballs, white smooth cylinders, and two different sized artificial plants. Test variables included four different detection heights (0.25, 0.5, 0.75 and 1.0 m) above each object, five sensor travel speeds (1.6, 2.4, 3.2, 4.0 and 4.8 km h-1), and 8 to 15 horizontal distances to the sensor ranging from 0 to 3.5 m. Horizontal distances of the objects to the laser sensor influenced the accuracy significantly, and the influence became weaker when the objects were closer to each other. The color and finish of object surfaces did not apparently affect the sensor detection accuracy. The average root mean square error (RMSE) and coefficient of variation (CV) in the laser sensor travel direction and object height direction varied slightly with detection heights, travel speeds, and object positions. The highest RMSE and CV were 83 mm and 50.9% in the horizontal direction, 41 mm and 15.2% in the travel direction, and 16 mm and 14.0% in the height direction, respectively. These tests demonstrated the potential adaptation of the laser scanning sensor and the specially-designed algorithm to measure complex-shaped targets for future development of automatic spray systems to achieve variable-rate functions in greenhouse applications.