Індикаційні стресові індекси технологічного характеру для ріпаку озимого

Abstract

Topicality. Precision farming technologies, which are becoming commonplace in modern crop production, require accurate and up-to-date information on the state of the plant vegetation. Technological stresses are challenge in the cultivation of winter crops, in particular rape, in industrial production using agrochemicals and plant protection products. Effective resuscitation measures are possible only in the initial stages of the growing season, and therefore require reliable and accessible information about the state of crops. Technical monitoring solutions were developed in the last century for satellites. Vegetation indices were created on the basis of the use of the soil line and were, for the most part, indicative of the condition of crops. For growers, it is important not only to assess the prospects of the harvest, but to identify the cause of stress in a timely manner to make a decision on technological intervention. The complexity of this is due to the interaction of factors of plant growth and development, described by the law of Mitcherlich-Tineman-Baule. The mentioned law concerned the combined action of natural factors, but can be applied to the actions of stressors of various kinds. The ability of a plant to withstand a stress factor - immunity - depends primarily on the provision of its nutrients and elements, because they form the potential of physiological and biochemical processes. This is one of the key points in identifying the nature of stress, which is especially important for technological stresses, in particular the phytotoxic effects of herbicides and other chemical compounds. Sheet diagnostics is effective for ground-based research to identify the nature of stress. However, such studies are weakly scaled for large production areas. The ability to determine high-resolution images of soil and plant parts obtained using unmanned aerial vehicles in images will allow the adaptation of leaf diagnostics technologies for large-scale use under conditions of competitive cost of research. Standard vegetation indices developed for satellite technologies are not suitable for these needs. Practical solutions require special stress indices that will level the subjectivity and allow to identify problem crops with high reliability. Processing of digital photography results in jpeg format was performed in MathCad. Research results and their discussion. Visual inspection of the experimental sites revealed: • the difference in size between healthy plants and affected, due to technological stress; • coloring the lower leaves of rapeseed in red and yellow and intermediate colors. The most promising for the selectivity to determine the stress of a technological nature is the fixation of abnormal color of the leaves. Samples of red, yellow and transitional leaves, as well as green upper leaves of affected and undamaged plants were subjected to spectral analysis. Soil samples with plant remains were analyzed separately. The obtained results show that the red channel in the optical range is the most suitable for the identification of abnormal color of rapeseed leaves, which is an indicator of technological stress. Thus, for the red channel the green leaves belonged to the range 0-210, red 70-230, yellow 140-255, and the soil - up to 90-180. That is, the yellow leaves can be identified quite accurately, because in the compatible range of 150-160 red channel was recorded up to 4% of the pixels of green leaves and up to 6% of yellow. Determination of the red leaf on this principle was unsuccessful due to a significant coincidence of spectral indices under the use of monochrome images. For the most suitable red channel at values of intensity of a component of color on the range of 100-140 it is necessary about 30% and 8% of pixels of green and red leaves accordingly. To reliably identify the yellow and red leaves, it is suggested to use an image of the object that takes into account several spectral channels simultaneously. Given the most common spectral channels of the optical range of commercially available industrial multispectral crop monitoring equipment (Slantrange and Parrot SEQUOIA), a combination of three RGB and two RG channels was considered separately. Among the proposed indices, the best sensitivity for the identification of red and yellow leaves belongs to the RLrg index. That is, the values of the index over 75 indicate the presence of technological stresses. The RLrgb index, with slightly poorer selectivity for the identification of affected leaves and soil, makes it possible to distinguish red and yellow leaves, which is promising for interpreting the causes of stress. In addition, for winter rape, it was found that the pixel distribution of green leaves in affected and healthy plants is in the range of 32 and 15 positive values of the RLrgb index, respectively. Since for healthy and affected plants the minimum value of RLrgb indices for green leaves was the same, this feature can be used as an adjunct to identify the technological stress of rape.