Testing the Suitability of a Terrestrial 2D LiDAR Scanner for Canopy Characterization of Greenhouse Tomato Crops

Jordi Llop, Emilio Gil, Jordi Llorens, Antonio Miranda-Fuentes, Montserrat Gallart
2016 Sensors  
15 Canopy characterization is a key factor to adjust pesticide dosage to the amount of vegetation. This 16 fact becomes especially important when the target is a fresh exportable vegetable like tomato 17 produced in greenhouse. The particularities of this crop, whose plants are thin, tall and planted in 18 pairs, make difficult their characterization with electronic methods. The present study attempts to 19 assess the accuracy of the terrestrial 2D LiDAR sensor for determining major canopy
more » ... eters 20 related to its volume and density and it establishes useful correlations between manual and 21 electronic parameters for leaf area estimation. The experiments were carried out at three different 22 commercial tomato greenhouses planted in a twin row system. The electronic characterization was 23 done with a LiDAR sensor (LMS-200, SICK) of 180º angle measurement by scanning the pair of 24 plants by both sides. The main parameters obtained were: canopy height, canopy width, canopy 25 volume and leaf area. From these, other important parameters were calculated, like the tree row 26 volume (TRV), the leaf wall area (LWA), the leaf area index (LAI) and leaf area density (LAD). A 27 general overview of the results show an overestimation of the parameters with manual 28 measurements due to the high definition of the profile obtained with this sensor. The estimation of 29 the canopy volume with the electronic device showed to be a reliable parameter to estimate the 30 canopy height, volume and density. Also, the LiDAR scanner demonstrated to be able to assess the 31 high variability of the canopy density along the row, resulting to be an important tool for canopy 32 maps generation. 33 Sensors 2016, 16, x FOR PEER REVIEW 2 of 4 achieve this goal is the dose adjustment to the real needs, avoiding overdosing and, therefore, 43 unnecessary PPP losses to the environment. 44 The greenhouse tomato crop, grown to be consumed as a fresh product, is very important in 45 Spain, with a cultivated area of 6189 ha [2]. Being important an accurate application of pesticides in 46 all kind of crops or circumstances, those related to fresh products to be directly commercialized in 47 the market, need a very accurate and safe use of pesticides in order to prevent health risks. Pesticide 48 residues on vegetables constitute a possible risk to consumers and have been a human health 49 concern [3]. In contrast, even though some works have been done at evaluating the optimal volumes 50 to be applied [4][5], not enough research has been done to relate all the parameters affecting the 51 relationship between the canopy characteristics and the amount of PPP according to the real needs. 52 Greenhouse tomato rises from the ground and develops a long stem, which is fixed by the 53 farmer to a fixed structure to make it stay in a vertical disposition. Therefore, this crop belongs to the 54 group of those called 3D crops, i.e., crops that present a complex geometry for the sprayer, in 55 opposition to the arable crops, that are treated as if they were a flat, 2D target. The constant pressure 56 that generates a constant liquid flow rate has shown to not to be valid for 3D crops [6], as the varying 57 geometry of the different individuals make very difficult to set a general application volume that 58 results in a satisfactory application quality. This fact led the researchers to set other systems that 59 focus on different parameters relative to the canopy structure. Thus, the first two methodologies to 60 appear were the Tree Row Volume (TRV) and the Leaf Wall Area (LWA). The TRV method consists 61 of calculating the canopy volume by assuming its prismatic shape, so the canopy height and width, 62 along with the row spacing, are the base parameters to determine the TRV, expressed in m 3 canopy 63 per ha ground [7-8]. The application volume will be proportional to this TRV parameter according 64 to a specific coefficient that will have different values according to the crop [9-11]. The Leaf Wall 65 Area, on the other hand, is based on the assumption that the canopy sides are completely flat, so they 66 form a "wall". The main parameter to calculate LWA is the canopy height [12], so this method 67 ignores the canopy width. The LWA is expressed in m 2 leaf wall area per ground ha. The sprayed 68 dose is calculated for every 10.000 m 2 LWA. These two systems are well implanted and nowadays 69 there is a general discussion between the countries of the European Union in which of these systems 70 should be used as the standard label dosing system for all the crops [13-14]. Nevertheless, in the last 71 years, different authors proposed alternative systems as the TRV and LWA do not take into account 72 a canopy parameter of major importance, the leaf density [14], so this method needs to be completed 73 with further information. Therefore, different dosing systems appeared for different crops, as 74 vineyards, citrus or fruit trees, like apple [6, 15-18]. Even though they differ in their basis, 75 assumptions and calculations, they all have something in common: they have to rely on an accurate 76 canopy characterization system. 77 Canopy characterization is a complex task that has been solved in the last years in very different 78 ways. The canopy characterization methods can be classified in two general categories: manual and 79 electronic methods. The manual methods are those that are based on manual measurements 80 performed with measuring tape, topographic milestone, etc. These methods vary according to the 81 canopy structure, and are much simpler in hedgerow orchards than in isolated trees or plants. Even 82 though they are reliable, fast and simple to use for the farmer, they become less useful for more 83 advanced task such as generating prescription maps for proportional spray application, like the one 84 proposed by the aforementioned dosing systems. In addition, the canopy density results extremely 85 difficult to evaluate with those methods, being necessary the complete defoliation of a representative 86 sample of plants to obtain reliable values. Therefore, the electronic methods seem to be a very 87 appropriate option to accomplish the requirements of the dose adjustment. Among the electronic 88 characterization methods, the more frequent are the ultrasonic sensors [18-20], the stereo vision [22], 89 the light sensors [23] and the LiDAR scanners [24-29]. According to Rosell and Sanz [30], LiDAR is 90 the most accurate technology to characterize the canopy, and in fact it showed to be very reliable at 91 predicting canopy parameters in different studies [20, 24, 31]. The LiDAR scanner is based on the 92 principle of Time-Of-Flight (TOF) to calculate distances, i.e., the sensor measures the elapsed time 93 between a laser beam emission and reception and automatically calculates the distance to the target 94 Sensors 2016, 16, x FOR PEER REVIEW 3 of 4 point [32]. This process is repeated along a plane in 2D scanners or in three dimensions, by rotating 95 the scanning plane, in 3D LiDAR. The 2D sensor is cheaper and can have a third coordinate by 96 moving it along the axis perpendicular to the scanning plane [24, 28], so it was more frequent for 97 canopy characterization. 98 The particularities of the tomato plants, which are thin, tall and planted in pairs, make difficult 99 their characterization with electronic methods, as it is difficult to identify the parameters related to 100 each individual plant. Furthermore, the narrow row spacing limits the field of view of the sensors 101 used. The aims of the present study are: 1) To assess the accuracy of the LiDAR sensor for 102 determining major canopy parameters related to its volume and density, 2) To establish useful 103 correlations between manual and electronic parameters for the leaf area estimation, 3) To take 104 advantage of the LiDAR technology to assess the variation of the canopy density throughout the 105 row, for being the basis to generate canopy density maps for pesticide dose adjustment. 107 428 Competitiveness (SAFESPRAY project AGL2010-22304-C04-04) and the European Regional 429
doi:10.3390/s16091435 pmid:27608025 pmcid:PMC5038713 fatcat:6perraizyzh7plpdrwy3zaww5u