The article compares the reproductive qualities of sows GGP, GP and P breeding levels in terms of industrial technology of pork production for different variants of their breeding. It is established that sows of maternal genotypes of Great White and Landrace breeds of Irish origin have a high level of reproductive qualities at all levels of the breeding pyramid in the industrial complex of the steppe zone of Ukraine. It is proved that sows of large white and landrace breeds in their purebred breeding gave birth to 32.1-35.2% more piglets, had higher by 35.9-37.5% fertility, less by 5.1- 6.4% share non-viable piglets in the nest at birth, increased by 11.4 -11.9% of the nest weight of piglets at birth and increased by 31.6% the number of piglets in the nest at weaning compared to their purebred counterparts of the synthetic terminal line Max Gro. At the same time, they were inferior to the latter in terms of high fertility by 22.0-29.2%, preservation by 5.9-6.8%, dew intensity of suckling piglets by 12.5-15.4% and as a result the weight of one piglet at weaning in 28 days by 14.2-18.8%. According to the complex of reproductive traits, sows of maternal genotypes were naturally better by 12.4-32.4% than their paternal counterparts. When comparing the reproductive qualities of sows of great white and Landrace breeds of GGP level (for purebred breeding) and their counterparts for GP level (for direct and reverse crossing), the advantages of GP level animals in the total number of born piglets by 2.1%, in fertility by 2.8%, by high fertility by 1.7%, by weight of nests of piglets at birth by 2.2%, by number of piglets at weaning by 4.1%, average weight of piglets at weaning by 1.3%, average weight of nests of piglets at weaning by 4.6%, the growth rate of piglets in the suckling period by 1.3%. At the same time, for the number of non-viable piglets and the safety of piglets before weaning, no significant difference was found between animals of these groups. A comprehensive assessment of the reproductive performance of GP sows using the SIVYAS index and the index with a limited number of traits showed the advantage of animals of this level over their counterparts with GGP level by 2.8-3.3%. It was found that local sows P level ♀VB × ♂L and ♀L × ♂ VB when inseminated with sperm boars of synthetic terminal line Max Gro predominated GP animals by 2.1% of the total number of piglets at birth, by 5.1% for high fertility, 2.3% -3.2% by number of piglets at weaning, 2.8% by weight of one head at weaning, 3.7% by weight of nest of piglets at weaning and 2.3% by growth rate piglets in the suckling period. At the same time, they were inferior to their GP counterparts by 2.7-3.3% in terms of the share of non-viable piglets and 0.6% in terms of fertility. A comprehensive assessment of the reproductive qualities of sows using the SIVYAS index and the index of reproductive qualities of sows with a limited number of traits did not reveal significant differences between sows P and GP levels. When comparing sows P and GGP levels (ma ternal form) found their advantages in the total number of piglets at birth by 3.9%, in fertility by 2.2%, in high fertility by 6.8%, in nest weight of piglets at birth by 6 , 3%, the safety of piglets before weaning by 1.7% -2.0%, the number of piglets weaned by 5.7% -6.5%, the weight of one head at weaning by 4.2%, live nest weight of piglets at weaning by 8.5%, the growth rate of piglets in the suckling period by 3.5%, but they have a 0.3% -2.7% lower proportion of non-viable piglets. According to a comprehensive assessment of sows using the SIVYAS index and the index of reproductive qualities of sows with a limited number of traits, the advantage of sows P level over GGP by 3.9% and 3.3%, respectively. P-level sows outperformed analogues of the Max Gro synthetic line in the total number of piglets at birth by 34.2%, in multiplicity by 59.1%, in nest weight of piglets at birth by 27.8%, in the number of piglets at weaning by 54.1%, by live weight of piglets' nests when weaned by 8.5%. But in the nests of sows of the synthetic line Max Gro found 4.2% higher share of non-viable piglets, 25.8% high fertility, 4.6%, survival of piglets before weaning, 11.8% weight of one head at weaning, 8.8% growth rate of piglets in the suckling period. According to a comprehensive assessment of sows using the SIVYAS index and the index of reproductive qualities of sows with a limited number of traits, the advantage of sows P level over GGP by 25.9 and 31.8%, respectively. Key words: reproductive qualities, preservation, multiplicity, nest weight, maternal lines, paternal lines.

No abstract

This paper considers a model of object detection on aerial photographs and video using a neural network in unmanned aerial systems. The development of artificial intelligence and computer vision systems for unmanned systems (drones, robots) requires the improvement of models for detecting and recognizing objects in images and video streams. The results of video and aerial photography in unmanned aircraft systems are processed by the operator manually but there are objective difficulties associated with the operator’s processing of a large number of videos and aerial photographs, so it is advisable to automate this process. Analysis of neural network models has revealed that the YOLOv5x model (USA) is most suitable, as a basic model, for performing the task of object detection on aerial photographs and video. The Microsoft COCO suite (USA) is used to train this model. This set contains more than 200,000 images across 80 categories. To improve the YOLOv5x model, the neural network was trained with a set of VisDrone 2021 images (China) with the choice of such optimal training parameters as the optimization algorithm SGD; the initial learning rate (step) of 0.0005; the number of epochs of 25. As a result, a new model of object detection on aerial photographs and videos with the proposed name VisDroneYOLOv5x was obtained. The effectiveness of the improved model was studied using aerial photographs and videos from the VisDrone 2021 set. To assess the effectiveness of the model, the following indicators were chosen as the main indicators: accuracy, sensitivity, the estimation of average accuracy. Using a convolutional neural network has made it possible to automate the process of object detection on aerial photographs and video in unmanned aerial systems.