Table 1 presents primary information about the dataset extricated from WoS containing papers related to DLA. The table gives an extended of curious data. For example, it take cognizance of the fact that the dataset is constituted mainly of articles (74.49% of all documents) published through 1051 sources (articles, book, letters, proceeding, editorial). The first publication dates from 2005.

Figure 1 presents the yearly scientific production of documents on DLA. According to this figure, the number of papers published on DLA in the WoS database is steadily increasing over years and the annual growth rate cumulate about 95.28%. This displays that the adoption of more focused agriculture is in a growing dynamic. Also, by the 27th of April 2022, the WoS database had already recorded 413 documents published on DLA. This represents 36.10% of previous year. So, the scientific production dynamic related to this field will be surely maintained.

The distribution of top 20 most relevant sources on DLA is displayed in Figure 2. It is clearly showing up that this shortlist is dominated by specialized journals on application of Information and Communication Technology (ICT) in agriculture. COMPUTERS AND ELECTRONICS IN AGRICULTURE is topping the list with 363 documents published on DLA. Seconded by REMOTE SENSING, with 169 documents. Thirded by IEEE ACCESS with 134 documents. SENSORS in the fourth place with 116 papers, and followed by FRONTIERS IN PLANT SCIENCE which has 77 publications. The remaining have between 58 and 15 publications dealing with DLA.

Figure 3 exposes the Source growth dynamics on DLA. Five journals are identified with the most relevant development based on the number of documents published, namely: COMPUTERS AND ELECTRONICS IN AGRICULTURE, REMOTE SENSING, IEEE ACCESS, SENSORS, FRONTIERS IN PLANT SCIENCE. Moreover, COMPUTERS AND ELECTRONICS IN AGRICULTURE experienced the most significant growth. All of these journals, except IEEE ACCESS, are related to application of ICT in Agriculture. This could explain why they are in these short lists.

In terms of the number of articles published, Table 2 displays the ranking of the 20 most productive authors on DLA as provided in the WoS database. It’s worth noting that each of the top 5 authors produced more than 60 publications in the following order: LI Y. (74 papers) leads the race, followed by ZHANG Y. (71 articles). WANG Y. (69 papers) and ZHANG J. occupied respectively the 3rd and 4th places. At the 5th place there is LI J. (64 papers). The remaining authors in this shortlist published between 57 and 36 papers. Most of the authors in this list are mainly from Asia, which shows that scientists in this region are very interested by DLA.

Table 3 displays the 20 most important affiliations on DLA papers distributed within the WoS database. The biggest number of articles goes to the China Agricultural University (1st position) with 403 records, taken after by Zhejiang University which has 191 reports (2nd position) and South China Agricultural University with 178 records (positioned 3^rd). The fourth position is held by Nanjing Agricultural University with 135 papers, taken after by Northwest AANDF University with 135 reports. It can be notice that as for author’s ranking 2 this list is also dominated by universities from Asia. This confirms that Asian scientists are very interested in DLA. 2. It also showed that few African institution appear in this list. This could be explained by the fact that, agriculture remains essentially traditional in this area.

Most globally cited documents are listed in Figure 4. It shows that the paper by Kamilaris A. (2018) has the most elevated add up to citations (929 citations) based on the WoS database. It’s taken after by Mohanty SP. (2016) with 774 citations. The third position is occupied by Kussul N. (2017) with 615 add up to citations within the WoS database. The paper by Ferentinos KP. (2018) is at the fourth position (485 citations), whereas Fuentes A. (2017) is positioned at the fifth position with 316 citations.

Table 4 displays the country’s scientific production on DLA distributed within the WoS database. China is taking the pace with 5858 reports, taken after by USA and India with 1697 and 756 articles respectively. At that point, we have JAPAN which is positioned fourth with 572 records, whereas South Korea is fifth with 560 reports. The remaining five nations of this confined list have less than 500 papers: Australia (467), Brazil (332), Pakistan (301), UK (283), and Germany (248). Tunisia (104) occupied the 28th place and is the most productive Africa country in term of number of documents published on DLA. In this category it is followed by SOUTH AFRICA (27 papers) and occupied 43rd place globally. The top 3 Africa countries is closed by KENYA (21 documents) which features at the 48th place in the general ranking.

Concerning citations, Table 5 presents the most cited countries. It shows that China, as the most productive country 4 in terms of number of papers published, has also the most elevated add up to citations number (11,229). It is taken after by the USA with 4555 add up to citations. All the remaining nations of this list have an add up to quotation check less than 2000 (between 1818 and 703). The first African countries in this list is TUNISIA (104 citations) which occupied the 28th rank globally. It is followed by EGYPT (35 citations, 44th global rank) and NIGERIA (11 citations, 56 global rank).

Figure 5 and Figure 6 present the most frequent words related to DLA according to the WoS database.

These figures reveal that the top 5 keywords used by authors in their articles are classification (27%), convolutional neural network (11%), model (7%), system (7%), agriculture (5%). This points out that research on Deep Learning in Agriculture turn around making a convolutional neural network model based system to accurately classify plant diseases. It could guide farmer to choose the most appropriate treatments in their crop management.

Other words in the list are segmentation with 140 occurrences, neural network with 139 occurrences. Finally, model, prediction and images with respectively 133, 128, 95 occurrences. These words are related to tools that scientists deal with in machine learning in general and particularly in deep learning.

Figure 7 presents some key collaborations between nations. The most substantial partnerships are initiated by China scholars, who establish links with their peers in different countries worldwide. On the other hand, African countries scarcely collaborate between them and on an international scale.