Automatic Wheat Ear Counting Using Thermal Imagery

Abstract

Ear density is one of the most important agronomical yield components in wheat. Ear counting is time-consuming and tedious as it is most often conducted manually in field conditions. Moreover, different sampling techniques are often used resulting in a lack of standard protocol, which may eventually affect inter-comparability of results. Thermal sensors capture crop canopy features with more contrast than RGB sensors for image segmentation and classification tasks. An automatic thermal ear counting system is proposed to count the number of ears using zenithal/nadir thermal images acquired from a moderately high resolution handheld thermal camera. Three experimental sites under different growing conditions in Spain were used on a set of 24 varieties of durum wheat for this study. The automatic pipeline system developed uses contrast enhancement and filter techniques to segment image regions detected as ears. The approach is based on the temperature differential between the ears and the rest of the canopy, given that ears usually have higher temperatures due to their lower transpiration rates. Thermal images were acquired, together with RGB images and in situ (i.e., directly in the plot) visual ear counting from the same plot segment for validation purposes. The relationship between the thermal counting values and the in situ visual counting was fairly weak (R2 = 0.40), which highlights the difficulties in estimating ear density from one single image-perspective. However, the results show that the automatic thermal ear counting system performed quite well in counting the ears that do appear in the thermal images, exhibiting high correlations with the manual image-based counts from both thermal and RGB images in the sub-plot validation ring (R2 = 0.75–0.84). Automatic ear counting also exhibited high correlation with the manual counting from thermal images when considering the complete image (R2 = 0.80). The results also show a high correlation between the thermal and the RGB manual counting using the validation ring (R2 = 0.83). Methodological requirements and potential limitations of the technique are discussed.

Automatic wheat ear counting using machine learning based on RGB UAV imagery

“…To date, automatic ear‐counting systems, regardless of the acquisition equipment, have been evaluated from the ground, using only a portion of the area of the plot (Cointault et al ., 2008; Zhu et al ., 2016; Sadeghi‐Tehran et al ., 2017; Velumani et al ., 2017; Zhou et al ., 2018a,b; Fernandez‐Gallego et al ., 2018a,b; Madec et al ., 2019; Fernandez‐Gallego et al ., 2019a, b). Although the use of a UAV platform allows for the acquisition of the complete area of the phenotyping micro‐plots, multispectral, thermal and laser sensors, with fairly low spatial resolution from aerial platforms, all remain relatively costly.…”

“…As an alternative to this approach, on‐ground automatic ear‐counting systems have been developed, based on RGB (red, green, blue), thermal, multispectral and laser images. In the case of thermal, multispectral and laser sensors, a few image‐processing techniques have been developed: for instance, color thermal maps and contrast‐limited adaptive histogram equalization (CLAHE) (Fernandez‐Gallego et al ., 2019a); threshold segmentation and de‐noising based on morphological filters (Zhou et al ., 2018a) for multispectral images; and in the case of laser sensors, voxel‐based tree detection and the mean‐shift approach (Velumani et al ., 2017). Nevertheless, RGB sensors have been widely used as proximal and remote‐sensing tools for many phenotyping tasks (Araus et al ., 2018) because of their relatively low cost (Qiu et al ., 2018; Araus et al ., 2018), high resolution (Deery et al ., 2014; Minervini et al ., 2015) and rapid adaptation to natural light conditions (Cointault et al ., 2008; Fernandez‐Gallego, et al ., 2019c), which allows RGB sensors to acquire a faithful representation of an original scene, even when mounted on aerial platforms with continuous and unforeseen movements.…”

High-Throughput Measurements of Stem Characteristics to Estimate Ear Density and Above-Ground Biomass

“…algorithms [17][18][19][20]. However, ear density estimation performances were generally limited to a comparison between the ears detected by the machine learning algorithm and those that can be visually identified by and operated on the image.…”

“…Further, the number of stems at relatively early stages may overestimate the actual stem density at harvest because of possible tiller regression as already pointed out. Previous scientists have used algorithms for estimating wheat ear density in-field conditions using RGB or thermal imagery [17][18][19][20]. However, these techniques, operated from the top of the canopy before harvest, may be limited when a significant number of ears are laying in the lower layers of the canopy.…”

Emerging thermal imaging techniques for seed quality evaluation: Principles and applications

“…Figure 4, thermal imaging systems of different designs have been successfully employed in different agricultural applications such as distinguishing soil surface crust, field nursery, detection of water stress in crops, yield forecasting, irrigation scheduling, identifying the ideal harvesting date, plant disease detection, fruit maturity evaluation, bruise detection, and monitoring of agricultural equipment(Danno et al, 1980;Fernandez-Gallego, Buchaillot, Aparicio Gutiérrez, Nieto- Taladriz, Araus & Kefauver, 2019;Kuzy et al, 2015;Li, Zhang & Huang, 2014;Manickavasagan, 2007;Meinlschmidt & Maergner, 2003;Roy et al, 2016;Varith, Hyde, Baritelle, Fellman & Sattabongkot, 2003;Villaseñor-Mora et al, 2017).…”