<< Volver atrás

Tesis:

Tecnologías de agricultura de precisión para la evaluación de poblaciones de Sorghum halepense (L) pers. en cultivos de maiz.

  • Autor: ANDUJAR SANCHEZ, Dionisio

  • Título: Tecnologías de agricultura de precisión para la evaluación de poblaciones de Sorghum halepense (L) pers. en cultivos de maiz.

  • Fecha: 2010

  • Materia: Sin materia definida

  • Escuela: E.T.S. DE INGENIEROS AGRONOMOS

  • Departamentos: PRODUCCION VEGETAL: BOTANICA Y PROTECCION VEGETAL

  • Acceso electrónico:

  • Director/a 1º: FERNANDEZ QUINTANILLA GALLASTEGUI, César
  • Director/a 2º: DORADO GOMEZ, José

  • Resumen: Sorghum halepense is one of the most serious weeds in maize in Spain. Due to its vegetative reproduction system, this species has a tendency to grow in compact patches. However, no information is available on the spatial characteristics of this weed. A survey was conducted to investigate the distribution patterns of S. halepense populations present in commercial maize fields in three major maize producing regions of Spain. A total of 232 ha were visually assessed from the cabin of a combine during harvesting using a three category ranking (high density, low density, no presence) and recording the geo-referenced data in a Tablet computer. On average, 4.7% and 11.4% of the surveyed area was infested with high and low density of S. halepense, respectively. Most of the infested area was concentrated in a few large patches with irregular shape. Management factors could explain much of the spatial distribution of this weed. Tillage direction was the main factor explaining patch shape: the length-to-width ratio of the patches was greater than two times in the tillage direction. In sprinkling irrigated fields, higher levels of infestation were generally observed close to the sprinkler lines. Areas close to the edges of the field had a higher risk of infestation that the areas in the middle of the fields: a linear relationship between distance from the edge and weed abundance was established. Because a few patches, located in some predictable parts of the field, represent most of the seriously infested area, site-specific treatments of these areas could reduce herbicide inputs until more reliable and precise detection and mapping systems become available. Visual mapping of S. halepense patches is a relatively simple and economic method to monitor the infestations of this weed at harvest time. In order to evaluate the reliability of this method to predict the location of S. halepense patches the following cropping season, prior to post-emergence herbicide treatment, three maize fields were surveyed at autumn and in the following spring. The comparison between harvest maps and spring maps did not show significant differences between the two maps. Therefore, characterizing the location of S. halepense patches from the combine appears to be a precise and cheap method for the construction of weed management maps to be used in the following season. With the aim of measuring weed spatial distribution automatically, a ground-based weed mapping system was developed. The system consisted of three different components: i) weed detection by optoelectronic sensors; ii) weed geo-positioning by DGPS receiver; and iii) a data acquisition and processing system. Three optoelectronic modules were mounted on the front of a tractor at 0.75 m intervals and at 0.60 m height above ground level. Consequently, the system was able to explore a 2.25 m band, corresponding to three crop rows, detecting the vegetation present in the middle of the inter-rows maize area. The weed mapping system was evaluated in three maize fields in the early spring (with the crop at the 4-to-6 leaves stage). The optoelectronic mapping system provided a good description of weed infestations, with 84% of correct measurements. The comparison among the results obtained with various simulated distances between sensors (from 1.5 to 6.0 m) indicated that a 4.5 m distance would be acceptable to construct accurate weed maps. Although visual weed cover estimations are a common method to quantify weed infestation level in crops and although this method has been used widely for weed mapping, there are some concerns regarding the reliability and consistency of visual estimation of weed cover. A study was conducted to assess the validity of this method, considering various factors which may influence the accuracy of human perception. Measurements were based on visual estimates of digital images acquired during three years in two maize fields. A total of 750 images were assessed independently by four experienced observers using three cover-abundance scales, varying the moment or the order in which the images were assessed. A sample of 95 representative images was used to validate visual estimations. While a good relationship was generally found between the visual estimates and some objective parameters (e.g., weed biomass), observers tended to overestimate weed cover at low weed densities and underestimate it when densities were high. Although weed cover assessment is a simple and relatively accurate method to quantify weed infestation level, our results showed that human perception was not reliable enough to determine actual weed cover close to the class boundaries both in the lower range (e.g., 5%, 10%) and in the middle range (e.g., 50%) of the scale and, consequently, can not be used in practice under threshold-based weed management programs.