Automatic Detection and Classification of Insects Using Hybrid FF-GWO-CNN Algorithm

Pest detection in agricultural cropfields is the most challenging task,so an effective pest detection technique is required to detect insects automatically.Image processing techniques are widely preferred in agricultural science because they offer multiple advantages like maximal crop protection,improved crop man-agement and productivity.On the other hand,developing the automatic pest mon-itoring system dramatically reduces the workforce and errors.Existing image processing approaches are limited due to the disadvantages like poor efficiency and less accuracy.Therefore,a successful image processing technique based on FF-GWO-CNN classification algorithm is introduced for effective pest monitor-ing and detection.The four-step image processing technique begins with image pre-processing,removing the insect image’s noise and sunlight illumination by utilizing an adaptive medianfilter.The insects’size and shape are identified using the Expectation Maximization Algorithm(EMA)based clustering technique,which involves not only clustering the data but also uncovering the correlations by visualizing the global shape of an image.Speeded up robust feature(SURF)method is employed to select the best possible image features.Eventually,the image with best features is classified by introducing a hybrid FF-GWO-CNN algorithm,which combines the benefits of Firefly(FF),Grey Wolf Optimization(GWO)and Convolutional Neural Network(CNN)classification algorithm for enhancing the classification accuracy.The entire work is executed in MATLAB simulation software.The test result reveals that the suggested technique has deliv-ered optimal performance with high accuracy of 97.5%,precision of 94%,recall of 92%and F-score value of 92%.

Haplotyping of Rice Genotypes Using Simple Sequence Repeat Markers Associated with Salt Tolerance

Salt stress is a major problem in most of the rice growing areas in the world. A major QTLSaltol associated with salt tolerance at the seedling stage has been mapped on chromosome 1 in rice.This study aimed to characterize the haplotype diversity at Saltol and additional QTLs associated withsalt tolerance. Salt tolerance at the seedling stage was assessed in 54 rice genotypes in the scale of 1to 9 score at EC = 10 dSm^-1 under controlled environmental conditions. Seven new breeding linesincluding three KMR3/O. rufipogon introgression lines showed similar salt tolerant ability as FL478 andcan be good sources of new genes/alleles for salt tolerance. Simple sequence repeat （SSR） markerRM289 showed only two alleles and RM8094 showed seven alleles. Polymorphic information contentvalue varied from 0.55 for RM289 to 0.99 for RM8094 and RM493. Based on 14 SSR markers, the 54lines were clearly separated into two major clusters. Fourteen haplotypes were identified based onSaltol linked markers with FL478 as the reference. Alleles of RM8094 and RM3412 can discriminatebetween the salt tolerant and susceptible genotypes clearly and hence can be useful in marker-assistedselection at the seedling stage. Other markers RM10720 on chromosome 1 and RM149 and RM264 onchromosome 8 can also distinguish tolerant and susceptible lines but with lesser stringency.

Informative ISSR Markers Help Identify Genetically Distinct Accessions of Oryza rufipogon in Yield Improvement

Inter simple sequence repeat（ISSR） polymorphism was used to determine genetic diversity and phylogenetic relationships in 90 genotypes of wild and cultivated species of Oryza from different geographical regions of the world. In all the 17 primers used in ISSR-PCR, a total of 11 464 bands were amplified at 253 band positions/loci. The primer UBC-809 amplified the maximum bands（1 059） at 21 band positions. UBC-810 and UBC-835 amplified the minimum of 391 bands each at 7 and 14 band positions, respectively. The mean polymorphism information content ranged from 0.44 to 0.84 and resolving power ranged from 8.69 to 23.53. Un-weighted pair group method with arithmetic mean dendrogram and population structure based on the 17 primers separated all genotypes into 4 major clusters with a genetic similarity of 53%–100%. The first two clusters consisted of 30 O. rufipogon accessions each. In the third cluster, O. nivara and O. longistaminata grouped as one sub-cluster and all other O. nivara accessions and cultivars grouped as another sub-cluster. The fourth cluster had only five O. rufipogon accessions which can be a source of new genes. Four sub-populations were identified within O. rufipogon and two sub-populations within O. nivara at K = 7. A subset of six primers with high resolving power values were the most informative and grouped all genotypes almost similarly as the 17 primers did. Use of these six highly informative primers in ISSR-PCR is a cost effective and robust method for assessing genetic diversity in large germplasm collections of wild rice species.

Yield Traits and Associated Marker Segregation in Elite Introgression Lines Derived from O. sativa × O. nivara

Introgression lines(ILs)derived from interspecific crosses are a source of new genetic variability.A total of 55 ILs derived from two crosses Swarna×O.nivara IRGC81848(population A)and Swarna×O.nivara IRGC81832(population B)were characterized for yield and yield-related traits/QTLs.Segregation of 103 simple sequence repeat(SSR)markers associated with yield-related QTLs was studied.Population A showed an average of 12.6%homozygous O.nivara alleles and population B showed 10.6%.Interestingly,three SSR markers,RM223,RM128 and RM517,showed conspicuous pattern of segregation.The distribution of parental alleles at three loci RM223,RM128 and RM517 linked to yield-related traits was unique.These markers flanked to several yield-related QTLs.RM223,flanking to qyld8.3,was heterozygous in almost all the 55 ILs except in IL10-3S and IL131S.RM128 on chromosome 1 and RM517 on chromosome 3 were mutually exclusive in 46 out of 55 ILs.These 46 ILs showed either of the marker allele RM128 or RM517 from O.nivara but not both.IL166S had both RM128and RM517 from O.nivara and the other ILs showed homozygous Swarna allele at RM517 except IL65S.Population structure assigned the 55 ILs to three sub-populations based on their genomic diversity.IL65S,IL166S,IL248S,IL7K and IL250K showed high yields in multi-location trials,and IL248S was released for cultivation as DRRDhan 40.