**4.1 Spectral pre-processing**

60 Agricultural Science

Fourier Hybrid Fig. 13. Visual comparison between Fourier and Hybrid approach for two different images.

Visual comparisons with hybrid space results show that high pass Fourier filtering approach eliminate more non wheat objects within the images. Fourier approach separates more efficiently ear groups. Calculation time for Fourier approach is few sec per image while its

High pass Fourier filtering gives global satisfying results. Although a close range of settings has been determined, inverse FFT remains a parameter that has to be adjusted according to input image. An empirical value has been found and gives good results for most of images but it could be optimized with an automatic threshold selection such as k-means methods

In the context of wheat detection, it has been observed that some ear objects are eliminated after cleaning step. These non-detections mainly correspond to near ground ears or ears massively hidden under leaves, hence, it should be relativized as too low ears may have development problems and may not be considerate within wheat yield. Ears that are located in over or under exposed part of image are not well detected but it is not due to the algorithm but to the quality of the acquisition, which is limited by the natural conditions. Small amount of very big leaves also remain after cleaning step and eliminate these artifacts

In the context of wheat ear counting, it is observed that counting error percentage decrease with number of ears in images, hence, best results may be obtained with images representing more area and yet, more ears. Actually, worst error, 5,56%, is obtained with

It is important to note that in most cases, Fourier approach returns slightly higher counts than manually counts. It should be due to missing detection, such as remaining leaves or over exposed part of images. Counts should be more precise with the including of shape analysis in cleaning step. In all cases, whatever the method used, the only way to obtain a

constitute a further axis of development, including shape analysis in cleaning step.

few minutes for hybrid space, with the same operating system.

(MacQueen, 1967).

only 36 wheat ears in the image.

right detection is to use 3-Dimensional information.

The images acquired by the camera (after a first internal processing taking into account the spectral sensitivity of the sensor) are images of luminance, which by definition depend on both the reflectance of the scene and lighting conditions. This is why a reference surface (gray ceramic), for which the reflectance Rc was measured in the laboratory, was systematically placed in each scene (figure 14). The average luminance Lc observed in the image to the reference can then be corrected reflectance for each pixel of vegetation :

$$\text{Rf} \equiv \text{Rc} \, \text{\* (Lf/Lc)} \tag{11}$$

where Lf is the observed luminance for that pixel.

It is important to note that the Rf value obtained is a "apparent" reflectance. Other phenomena are to be considered, namely:


Texture, Color and Frequential Proxy-Detection

left)

**4.3 Results** 

Image Processing for Crop Characterization in a Context of Precision Agriculture 63

Fig. 15. Calibration image and sample positions (the reference ceramic can be seen on the

Figure 16 shows the image test and the classification results. We observe an excellent discrimination between wheat and weeds whatever the local conditions of illumination (shadow), with the exception of one type of weed (red circle) which was not present in the

Fig. 16. Test image and classification results (light gray: ground, red: weed, green: wheat)

These results show the very high potential of hyperspectral image processing to extract

**5. Proxy-detection image processing for seed development and germination** 

Several aspects of seed quality are tested routinely to minimize the risk of sowing seedlots that do not have the capacity to produce the desired crop. Amongst these tests, seed germination is important since it represents the percentage of pure seeds that have the potential to produce established seedlings in the field. The rate of germination corresponds to the reciprocal of the time needed for a given germination percentage to be reached (Halmer, 2008). Accurate procedure of germination tests performed in laboratory is defined

**performance of chicory achenes by chlorophyll fluorescence** 

calibration image, and was therefore not included in the model.

pertinent information for agronomic applications.

by the International Seed Testing Association (ISTA).

These two phenomena are taken into account by applying a standard-centering operation (SNV, or "standard normal variate") on each spectrum of the image (Vigneau et al., 2011).
